Gemini 4 Neon

(blog.google)

652 points | by yigitdemirag 8 hours ago ago

557 comments

  • HJain13 2 hours ago

    Edited: For formatting and from feedback

    So if I am getting this right?

      +-------------------------------+--------------+--------+--------+----------+-------+
      |              Lab              |      XL      |   L    |   M+   |    M     | S/XS  |
      +-------------------------------+--------------+--------+--------+----------+-------+
      | Google (Gemini 4 Family #1)   |              | Argon  |        |          |       |
      | Google (Gemini 4 Family #2)   |              | Carbon | Neon   | Hydrogen |       |
      | Anthropic (Claude 5.2 Family) | Fable/Mythos |        |        |          |       |
      | Anthropic (Claude 5.5 Family) |              | Opus   |        | Sonnet   | Haiku |
      | OpenAI (GPT 6.2 Family)       | Astra        | Sol    |        |          |       |
      | OpenAI (GPT 6.1 Family)       |              |        |        | Terra    |       |
      | OpenAI (GPT 6 Family)         |              |        |        |          | Luna  |
      +-------------------------------+--------------+--------+--------+----------+-------+
    
    
    
    Notes:

      * Gemini 4 Family
        - Argon   -- ++Train --> Carbon
        - Carbon  -- Distill --> Neon, Hydrogen
      * Gemini 3 Family
        - 3 Pro   -- ++Train --> 3.1 Pro
        - 3 Pro   -- Distill --> 3 Flash
        - 3 Flash -- ++Train --> 3.5 Flash ... -- ++Train --> ... 3.8 Flash
      * "S/XS" classification is due to weirdly low pricing for models by Ant/OAI and not by raw perf
    • riskassessment an hour ago

      I'm not sure it even makes sense to align models on price given that Neon appears to be considerably better than Sonnet/Terra. Not to mention that there are rumors of a flash-lite type model even faster than Hydrogen.

    • primaprashant an hour ago

      As I understand it, Gemini 4 family should be shifted 1 step to the left. Argon / Carbon are supposed to be XL and in the same class as Fable / Astra.

    • HJain13 an hour ago

      I have based this on model pricing (which I guess also corresponds to their labs's idea of performance/how they compete) and for google their model cards which specify which model is derived from what

    • satvikpendem an hour ago

      This is illegible on mobile.

      • wwizo an hour ago

        Rotating screen 90 degrees seems to fix it

      • smcameron an hour ago

        Mobile sucks that way.

  • impulser_ 8 hours ago

    Is there any theory on what is Anthropic's secret sauce? I have been using the latest Gemini models and they just don't seem as good as Opus 5.5 despite the benchmarks. Yeah it comparable on some tasks maybe slightly better but as an overall general intelligent model nothing compared to Opus 5.5 and it hard for me to understand why. Opus feels like human level intelligent like I'm talking to a coworker basically while GPT and Gemini feel like I'm talking to an AI and have to change how I talk. It hard to explain and maybe I'm the only one that feels this.

    Is it because Anthropic put a ton more effort into the character of the model that it basically roleplays as a human and that why it tends to think better when giving tasks that it might not be trained heavily on.

    I have used Carbon more than most model to try and get the same results but it just feels like i'm talking to a smart AI. Which is fine if you are asking questions, researching an idea, but if you actually want to work on something, build something, it just falls apart. Like Opus can run for days on a task and it feels like it understand what it was doing 2 days ago even when it might not be in it context while other models feel like they breakdown and get lost or falter as the task grows.

    • svara 7 hours ago

      I think Gemini 4 (Argon is the one I've been using) is very clearly better at writing English than any Anthropic model. It's not better otherwise but in that the difference is vast.

      To the point that I now use a workflow of delegating from Opus or Fable to Gemini4 to write commit messages and PR descriptions, it mostly solves the Claudese problem.

      • hikarudo 6 hours ago

        I just use Opus with the simple-english skill. No more Claudese.

        • NikhilVerma an hour ago

          I've been maintaining SlopSift - https://slopsift.dev/ for a while it's powered by a dependency parser to deterministically lint English. Then I wrap it around Claude either as a skill or a hook. It's really simplified the AI speak for me

      • behnamoh 7 hours ago

        > To the point that I now use a workflow of delegating from Opus or Fable to Gemini4 to write commit messages and PR descriptions, it mostly solves the Claudese problem.

        how do you do this w/o breaking bank? for Argon to write good commit messages, it must read the diffs and all the work Opus has done, which could easily eat up its usage quota.

        • lilbigdoot 6 hours ago

          I use a $20 a month plan with Argon and I honestly find it pretty usable for asking a lot of questions if I don't keep it all in one session, or it can do a few small coding tasks a day for me until I hit my limit. I haven't hit any limits with the smaller models cuz I don't like using a lot of LLM code in my personal projects since me learning a new thing or area by coding is one of the main motivations. Generating messages from diffs/notes after the work is done is pretty cheap I imagine.

        • rolosa 7 hours ago

          Can you just have opus hand off the changelog to Gemini and have it write from that?

        • spolitry 6 hours ago

          You vastly underestimate size of context the model can process.

      • disillusioned 4 hours ago

        Gemini has reliably been the best writer of the bunch, and Argon has blown me away with how "human" it can sound. It has so much less of the AI smell than the others, though Opus 5.5 definitely lobotomized the absolute sloppery that was Opus 5, at least.

        • antupis 3 hours ago

          Yeah usually I use Gemini for writing. Anthropic and OpenAI models even with heavy prompting sound like LLMs where Gemini is way more natural sounding.

    • ehsankia 8 hours ago

      Every day I'm just more and more amazed at Opus 5.5 and how ahead it was. Even with the insane pace we're going at, it's amazing that two months later, we're barely putting out models that hit the price/speed/quality of Opus 5.5

      No wonder they went for the 5.5 naming, instead of the .1 increase. From just the benchmarks, it seems like neon is basically on par with Opus 5.5 which came out 2 months ago.

      • aenis 8 hours ago

        Same here. Spin up a few long horizon tasks, have it manage the whole thing as a project, you can talk to it, interrupt it, engage in side-quests, its like talking to a human.

        Have another session on another machine (say, IOS build on a mac, and server build on a linux?). No problem, they will happily coordinate with each other through the /remote-control protocol, and then you feel like you are in a crowd, with those sessions talking to each other and over you.

        No such feel with OpenAI nor Google models for me. They are absolutely great for lots of things, but nothing comes close to the feel of Opus 5.5 and Fable.

        (I am biased; i mainly use the anthropic models so i got used to them).

        • jwpapi 8 hours ago

          I dont even feel like fable 5.2 is better. Imho the best model rn is Opus 5.5 MEDIUM.

          • p1necone 4 hours ago

            The first time I tried to use fable 5.2 it made a bunch of really amateurish technical design assumptions about my project that cursory reading of docs would have disproven (but also should have just been discarded because of taste reasons anyway). I switched back to Opus 5.5 after that.

      • pjankiewicz 8 hours ago

        I agree Opus 5.5 is something different entirely. It is even hard to find tasks for other agents to be honest where you know the results will be picked apart when Opus 5.5 reviews them. In my setup Opus 5.5 uses codex from time to time when I'm running out of credits and it decided to run 2 astra agents for the same task to improve its performance.

      • nicoburns 7 hours ago

        > two months later

        Mad that this is seen as a long time horizon!

    • foohis 8 hours ago

      Anthropic spent a lot more time editing their pre-training data with higher-quality material. For example, they used actual books for pre-training, because books are generally run through human editors for higher quality. (this is from the copyright lawsuits)

      Other companies would use completely raw internet data for pre-training, which is basically trash. Anthropic's secret sauce is that they filtered and edited down their data for only the highest-quality material. You really don't need quadrillions of tokens to pre-train, as that's just random noise. A smaller-but-higher-quality data set is better in every way.

      Meanwhile its obvious that companies like Google would use trash data like email and chat logs or Internet forums for pre-training. Just complete trash data.

      This is a classic case of "less is more" and "garbage-in-equals-garbage-out"

      • hackingonempty 7 hours ago

          Anthropic spent a lot more time editing their pre-training data with higher-quality material. For example, they used actual books for pre-training, because books are generally run through human editors for higher quality. (this is from the copyright lawsuits)
          [ ... ]
          Meanwhile its obvious that companies like Google would use trash data like email and chat logs or Internet forums for pre-training. Just complete trash data.
        
        Google has been scanning every book they can get their hands on for 20 years or more. They have a bunch of patents on non-destructive book scanning gear. Surely they are using that data to train their models.
        • foohis 6 hours ago

          won't help if they also include trash internet text in the training set as well.

          • etoxin 5 hours ago

            it's kinda known that the gemini models write better english which goes against the "trash data" remarks.

        • wahnfrieden 6 hours ago

          Surely they are destroying scanned books for training

          • mastax 3 hours ago

            They invented those nondestructive scanners before the court cases that effectively mandated destruction.

            • wahnfrieden 2 hours ago

              Why does that allow them to use the scans for LLM training? Those scans were used for a different purpose. I don’t understand the logic.

      • impulser_ 7 hours ago

        I don't think it pre-training, because surely these labs all have similar data and Google should be very good at ingesting, cleaning, and filtering data.

        I honestly think it just their post-training is fantastic and I think it has something to do with them being very opinionated on how they want the model to act that it basically the best at roleplaying a human. Humans are very good at completely and figuring out things that there aren't "trained" on if given basic set of intelligence. So when it working on these long tasks that might not be directly in it training data it still reason through it very well.

        They got made fun of for this in the pass because they do a lot of research on this including hiring a philosopher.

        This is just my best guess, because you don't see any other labs talk about how they made the model act and Anthropic is basically obsessed with it and talks about it every model release.

        I think they just have better "taste" and that taste is their secret sauce.

      • nl 7 hours ago

        This is incorrect.

        Everyone uses books as well as internet data.

        Anthropics difference is they have a huge data budget for buying RL environments for real world tasks.

        • positr0n 4 hours ago

          I get recruiter emails for $150/hr to do code review training for anthropic (W-2, although short duration contract).

          • nl an hour ago

            Yes, and if you look on LinkedIn there are many companies doing similar tasks in areas from accounting to physics to HR. They sell to all companies, but Anthropic seems the most aggressive. This from last year:

            > According to The Information, leaders at Anthropic have discussed spending more than $1 billion on RL environments over the next year.

            https://techcrunch.com/2025/09/21/silicon-valley-bets-big-on...

      • mlmonkey 6 hours ago

        Microsoft (MSAI) showed that quality >> quantity with their Phi model a couple of years ago. It was a small model, but trained on very high quality data, and did better than much larger models.

      • nikhilsimha 38 minutes ago

        i saw a researcher at uw claim that claude is trained on genome sequences and that others arent.

        genome sequences are probably very dense in “reasoning”

        • oasisbob 19 minutes ago

          I have a weak theory that some of the strange contortions in Claude come from training on scientific literature.

      • potamic an hour ago

        Is this your opinion or you have a source for your claim?

      • amarcheschi 5 hours ago

        I can't say what precisely, but other big companies are also doing what they can to get human data of high quality, just look for all the companies that popped like mushrooms that do ai reinforcement tasks, micro1, ethos, mercor ecc

      • mrheosuper 4 hours ago

        Is that reason why Claude talks so like .... a nerd.

      • olalonde 5 hours ago

        Where do you get that information from? Or is it just speculation?

      • quixoticelixer- 8 hours ago

        That doesn't really make sense because as good as Opus 5.5 is it writes awful terrible, disgusting slop.

        • weitendorf 7 hours ago

          Anthropic doesn't want people to use their models for the chat/persona style use cases OpenAI has spent more time and effort optimizing for.

          If you look back at GPT2 and the early days of AI safety/risk this was actually something the kinds of AI researchers that joined Anthropic didn't want - RL on coding->using that accelerate development and to bootstrap capabilities across domains was always the plan, not making it easier for people to spam/fake homework or trap people in problematic human-AI attention/relationship tarpits.

          I personally find that Opus 5.5 and Fable 5+ are generally fine to chat with but just very wordy and culturally "LessWrong/essay-guy". Which, if a model is going to have a personality, is close to the one that I want (though it does piss me off when it gets preachy about standards Anthropic clearly doesn't hold themselves against)

          I also want to immediately recognize AI generated content on the web or from another person, so I'm very glad it has a distinct voice

          • quixoticelixer- 7 hours ago

            Yeah maybe that's the case, and maybe I am more sensitive than other people. But I don't really use AI for that sort of thing but even for technical stuff I still find it's writing style a real turn off. Gemini still has lots of AIisms but it's a lot more tolerable.

        • LeoPanthera 7 hours ago

          If the training set is smaller, doesn't that make sense though? There will be much less variety in its style and so its going to make very recognizable output.

          • quixoticelixer- 7 hours ago

            Yeah that could be it. I guess the thing about what makes AI writing so repulsive sometimes is not that it's bad like a human would be bad, but that it does the same things over and over.

        • foohis 7 hours ago

          It's a much, much better writer than literally every other model.

          • quixoticelixer- 5 hours ago

            It's better than Opus 5 but it still has that incredibly punchy contrasty LLM style that worse than most other models

            • thriving_earthw 3 hours ago

              I sometimes use Gemini for quick searches or stuff and the writing style is such a breath of fresh air. I mean I've been so locked in to the anthropic ecosystem because I find that code, the desktop app, and the ability of their models to understand what I mean and execute so well is incredible. But sometimes I'll read something it writes and I'll want to throw my computer at the wall. I just become apoplectic with rage for a second before I of course cool down and move on.

              • BigGreenJorts 3 hours ago

                I've been feeling this more and more lately as I've been trying to delegate technical writing to it. It's so horrifically repetitive, verbose and somehow non specific I can't even really use it as a base to edit down to something legible.

      • UltraSane 7 hours ago

        I think by now everyone training trillion parameter models is using almost every book, magazine, research paper, etc they can get.

      • Topfi 7 hours ago

        Purely outside looking in, but yes, I'd argue you are right and higher-quality is a major part of it, both for pre- and post-training. There appears to be a far greater willingness to adopt specifically tailored training sets from third-parties, consistently update and build on them.

        From simple frontend work to strategically adding virtual rulers when working with LIDAR scans to extract more accurate data, there are very clear points when Anthropic introduces a new external data set into their mix and while I have seen the same from competitors such as Google, there simply appears to be a greater tendency to stick with these once they have them.

        As an example, at release it was quite apparent that Gemini 3 Pro, 3 Flash and Opus 4.5 had a similar, possible even the same frontend-focused training base bought in from a third party. But while Anthropic has had both smaller and larger improvements across each subsequent Opus release, I have not noticed such a consistent improvement with Gemini models over the last year. Maybe Anthropic simply buys updates more often, maybe they create more artificial PRs on top, who knows.

        Same with the LIDAR scans, there was a point a few months ago when models started to exhibit a very systematic way of handling visual references in some few leather-crafting tasks, enabling whole new ways of working more efficiently combining existing measurements with images. Step-by-step pixel measurements, compensation for angles and poor captures, etc. Have since seen this in other models, again very much feels like a bought in data set that they are pushing as far as it goes.

        Then there is 3D capability with models like Muse Wonder Full. I have not used it myself, if it ain't on API, it doesn't exist, but have seen PRs and do find the model output very parsable, despite it being terrible at any modelling task because these capabilities do not emerge from just training on vast amounts of data, they require high-quality, well tailored, step-by-step reference data of the entire process.

        Harness wise, while pure output performance is very consistent across harnesses in my evals (I nowadays use both OpenCode via API and the labs own harness three times each to normalise), Anthropic also appears to be far more willing to use the time between releases to prepare features for the next. Stuff like workflows, proper subagent implementation, advisor, etc. often have found there way into Claude Code either in-between models or were first seen with a model that couldn't fully leverage these. Also feel that explained the good overall launch experience Fable 5 delivered, they simply used the time between announcing Glasswing and launching full access to get the UX/DX polished.

        Meanwhile, I am still testing GPT-6.2 Astra and Sol so will refrain from making any more judgement on either till I have all data, but apparently there has been an extensive last minute rewriting of the system prompt to improve performance and "behaviour". Given the gap between GPT-6 Astra and yesterday, along with statements by OpenAI staff alluding that they had this model ready for a while, such things should happen in a far more structured, systematic, well thought out manner, not last second after some negative outside feedback. Heck, from what I know Anthropic, unlike any other lab, does zero early access outside the limited hardening for very exclusive partners like Glasswing. Certainly no early access for regular users, reviewers or the like from what I've heard, so Anthropic does all that prep-work for model releases with less feedback than other labs.

    • antupis 3 hours ago

      It depends for programming Opus/Fable is the best but eg for code reviews or documentation Carbon is better. I think here Google own code review/documentation practices and thus probably best dataset in whole world really pays dividends.

    • cavisne an hour ago

      My theory is firstly its focus - they got to skip to GPT 4 (ie clean room implementation by the same people), they never cared about consumer or video gen or image gen.

      Secondly, they somehow made all this work on public clouds with really crappy (compared to baremetal nvidia) networking. I think thats probably a lot of post training (which doesn't need such a large/tightly connected cluster), and some architecture that is less reliant on a tightly connected cluster for pretraining. This would also explain why they have been able to use Trainium, and somehow beat Google while using TPU's.

    • surround 8 hours ago

      I suspect that Anthropic has more user data to train on from Claude Code. Claude Code became popular sooner than Codex and Antigravity, and remains more popular - thus more data to collect.

      Cursor was popular before any of those three, and just look at how much Grok improved after they acquired Cursor's data.

      • behnamoh 7 hours ago

        > Cursor was popular before any of those three, and just look at how much Grok improved after they acquired Cursor's data.

        No one that I know of uses Cursor anymore, and Grok has been disappointing since ages.

    • porphyra 5 hours ago

      They just have an incredible data flywheel I think. Since they were first to market with claude code, they have more training data than others. They also pay a lot ($150 per hour) for SWE experts to train the AI. I get some random recruiting emails occasionally from companies affiliated with them.

      • jbonatakis 5 hours ago

        Yeah I’ve gotten the same thing. If anyone from Anthropic is reading this: it’s not enough to make me quit my job for a short term contract. Try doubling it and we’ll talk

    • m_ke 7 hours ago

      Anthropic has a proper feedback loop with claude code. They see way more real production data and can detect where their models are weak.

      So they can train on prod traces or at the very least use user prompts to seed good synthetic tasks to train on.

      • chilmers 7 hours ago

        OpenAI has plenty of feedback from Codex. Maybe not quite on the same volume as Claude Codex, but still massive. So I'm not sure that's the whole story.

        I doubt there's a single reason for why Anthropic are ahead right now. Building a frontier model requires executing well in a wide variety of things. It's probably a result of being slightly ahead in various aspects of of data collection, model architecture, pre and post training, harness development, etc.

        It's a shame that so much of what's happening is being closed doors. Maybe one day, decades from now, somebody (or some AI) will write a definitive history of this period, explaining what was going on at the various labs, how their research compared, and what the major players really thought.

    • thanhhaimai 8 hours ago

      It's likely that the agent configs you have might overfit to Claude models.

      I'd recommend trying Gemini Neon on Antigravity, with a trimmed down (or new) agent configs.

    • irthomasthomas 6 hours ago

      This model matches Fable 5.2 for about 40% the cost. Not sure how you can take from this that anthropic has some secret sauce?

    • thatxliner 4 hours ago

      > Is there any theory on what is Anthropic's secret sauce?

      I'm planning to write a blog post on it, but my theory is that it's aligned to not be "politically correct" but instead to have a solid stance. In the Claude Constitution, they explicitly say how they train it to have judgement and not "follow rules"

    • jorl17 8 hours ago

      Here's my average experience with Carbon.

      (After a long conversation with existing messages)

      User Prompt: Ok, so SSH to the unraid server and update the plex db

      Reply:

      I need to think how to answer this.

      The user is asking me to SSH into the unraid server. This seems to imply there is a machine to which we can SSH to. I should look for places where hosts may exist.

      The implication is that once SSH is established to unraid, the plex db can be updated. It seems plex is a service that the machine has in which we can run updates.

      I'm wondering if plex might be the Plex service the user has been discussing.

      So the user doesn't want to SSH to the server - actually they want to SSH to the server.

      ... Formulating a plan ...

      [some messages later]

      Aha! THIS IS THE SMOKING GUN! `~/.ssh/config` HAS THE REQUESTED HOST!!! IT'S TIME TO CONNECT!!

      [some messages later]

      Ok, time to act. I'll ssh to the server and follow coherent rules.

      ----

      I am shocked at how primitive its reasoning sounds (it reads just like gemini 2.5-pro's old reasoning). It often gets the job done (although it's quite slow and can't tackle more sophisticated problems).

      Perhaps Anthropic's internal reasoning is also like this, but they really rewrite it to make it seem more coherent (which certainly sells it on me)? I almost want to click stop just by reading this nonsense.

      • iamdelirium 8 hours ago

        All displayed reasoning is not real reasoning for all models now. It's fake since they can be trained on.

        • jorl17 8 hours ago

          I want to believe you, and it's what makes the most sense! But if that's the case, why is google (intentionally?) presenting such terrible reasoning traces? To make sure that if they're trained on, they're trained on garbage? I guess it makes sense, but still...

          I was not really fabricating the example I gave too much. Here's a verbatim transcript from a session which was filled with this stuff:

              ... thought LET'S READ RecordReader.cs RIGHT NOW!! RecordReader.cs IS WHERE EVERY SINGLE TABLE, ReadPlan, DecideValueVisibility, ReadRowCounts , AND ShowValues READS ITS DATA!! Let's read RecordReader.cs (lines 1-300)!
          • 8note 7 hours ago

            reasoning traces also arent real reaaoning. its what tokens warm up the outpt the best

          • dist-epoch 7 hours ago

            What you're seeing is a small model summarizing and possibly obfuscating the real thinking.

            • jorl17 7 hours ago

              Yes, I understand it.

              It seems rather ironic that Anthropic givens the most logical reasoning traces, though. You'd think if they're ahead they'd either hide them or make them produce garbage (like what seems to come out of Google's models).

              • orbital-decay 7 hours ago

                Neither has any relation to the actual reasoning trace. The summary is actively misleading and you'd be surprised how different it is from the raw CoT of the model. Just ignore it.

                If you want to see example traces without jailbreaking claude or gpt, see open models that display the CoT without trying to hide it. It's pretty useful sometimes, with usual caveats (it deceptively looks like English but is heavily reward hacked, and it's generally a weak proxy for model's reasoning process).

                • jorl17 7 hours ago

                  This makes sense. Nonetheless, I find that reading Anthropic's reasoning traces is sometimes a trigger for researching more stuff. Often I'll see models thinking (or fake thinking, whatever) about something and I'll be curious to check out more about that topic if I'm not particularly familiar with it.

                • deaux 6 hours ago

                  Off-topic, but this kind of conversation tells me just how little the average HNer still uses LLMs outside of specific coding tasks, or at least pays attention to them. Which in turn explains so many comments. Probably a good thing though.

                  You're of course completely right. Interestingly the "summaries" have only become more diverged over time. I now believe the big 3 are barely summarizing at all any more. I deem it likely that the small models don't get the CoT at all, they just completely make up something "plausible" solely based on the user message with a little bit of previous context. Much of the time it will be the complete opposite of something I asked for, or barely related.

                  Anthropic has changed this up multiple times over the last 12m, I'm talking about the iteration that was around the longest this year. I think they changed it again within the last feww weeks and potentially started actually summarizing again.

                  Your suggestion is the right one. Everyone who uses LLMs for their job should spend some time with open models like Deepseek and see what the thinking actually looks like.

        • behnamoh 7 hours ago

          Not only that, Anthropic injects fakes reasoning and tool calls in there to make it harder for distillers. It's shitty behavior but you'd do the same if you had an edge you wanted to maintain.

      • aenis 8 hours ago

        One of my long horizon tasks benchmarks involves writing logic controlling a team of soldiers in a simulated environment. Argon and Carbon both made such discoveries such as "it would seem that i need to be pointing at an enemy to shoot them". Its almost entertaining to read, its like its discovering the world on first principles.

        (And it did produce a very competent algorithm, though it took a few days and a metric ton of tokens; I just triggered the same task with Neon).

        • jorl17 8 hours ago

          I'm curious about how it goes with Neon! This night I'll leave Neon with some lofty goals on a project and I'll see how it fares too.

        • aatd86 6 hours ago

          lol Aenis the god.

    • jaccola 8 hours ago

      Claude annoys me with condescension sometimes... makes it feel far less like a 'tool' and can even come across as a bit judgemental!

      Just today I was filling out a security questionnaire and it said "I won’t write a PDF that claims zero credentials while this token is in the logs." But we were in the process of removing this from the logs so it was perfectly accurate to write and I wish it had deferred to my explicit request...

      I think Anthropic overcorrected on the sycophancy; Google models to me seem a nice balance of highlighting issues / risks while still doing what I actually ask.

      • aenis 8 hours ago

        Anthropic models won't lie. When we fix security issues identified with sth. like Wiz and are writing a report, it won't write a summary suggesting that something we just fixed wasn't broken (even if that thing never made it to production). It has something akin to moral standards :-)

        • willis936 7 hours ago

          I've seen them ignore clear directions that in documentation it's supposed to read after a compaction. That seems a lot like lying to me.

      • enraged_camel 8 hours ago

        Claude models will go to extraordinary lengths to ensure sensitive data does not end up in their context window. It's a bit nuts. But frankly I appreciate it a lot. Even when it is annoying I know the model is doing the right thing.

        • hadlock 7 hours ago

          Opus 4.x really struggled with secrets rotation in sops, even in 5.5 it's constantly emitting secrets to the shell history. Thankfully everything gets tested in dev, but Claude models are the reason why we have such good secret rotation now.

        • impulser_ 7 hours ago

          Yeah, I notice this as well. I was building something to help me understand logs from the different coding agents and it wrote redaction code to redact secrets out of the logs just in case if the logs had any so it wouldn't see them.

    • danduma 5 hours ago

      You are not alone. I have the exact same experience: opus just gets it and the others feel like cheap toys

    • epolanski 8 hours ago

      > Is there any theory on what is Anthropic's secret sauce?

      Lots of PR it seems?

      I have all of Ant/OAI/OpencodeGo subs and test lots of models and I just cannot understand how and why some people find specific models so ahead, when it seems to me like reality is so nuanced and so specific to the particular task and how you approach it.

      • deaux 6 hours ago

        More like, most people especially on here do 90% of their "tests" on very similar tasks - coding on specific codebases - and the other 10% is just asking stuff in chat. So they never get to experience the jagged edge. Anthropic models could be horrific at translation to Nahuatl or greenhouse cherry tomato management compared to Gemini and OpenAI but there's one person on HN doing such things and tens of thousands doing enterprise SaaS coding.

      • WarmWash 8 hours ago

        Never underestimate the power of minimalist hand drawn graphics on an off white background with a clean typeface.

    • neximo64 4 hours ago

      Exclusivity on RL environments they license.

    • e12e 5 hours ago

      How do you evaluate opus 5.5 as "better"? My experience has been that Argon was ok - maybe equivalent to opus - but carbon in orchestration mode was finally able to rewrite some early experimental (ai slop) front end code - and break a bit out of UX/UI rut that a couple of projects were stuck in.

      It might be that opus is better in better structured projects, and carbon is better at (re)structuring poorly structured projects?

      Carbon is a better project manager/product owner - opus a better programmer?

    • baq 8 hours ago

      Opus 5.5 leads the CanTellForestFromTheTreesBench 42.0.

    • sudohalt 4 hours ago

      There isn't because Opus 5.5 isn't better. There is no objective reason to believe Opus 5.5 is better.

    • atemerev 4 hours ago

      The secret sauce is sandwich statements. Just look at any Claude response. Positive-negative-positive. Bamm, and we are hooked.

    • risyachka 6 hours ago

      Results usually reflect company culture and with big companies no matter what you do - if the culture is not there - you are doomed.

      Google is the best example of this with their product graveyard. No matter what they release- they just can’t make a good/nice/usable product.

      Just check google ai studio as one of latest examples.

      • spolitry 6 hours ago

        AI studio isn't great, but what is a better lightweight studio product than AI studio?

  • declanjackson 8 hours ago

    Something that is interesting is how it's positioned in the overall fleet of Gemini models. Although overall more Intelligent than Carbon, Carbon is still positioned as the 'long horizon' model. This is clear on AA Index, where we see higher overall scores for Neon but lower scores in evals like CritPt which test research-scale reasoning capabilities.

    With the frontier getting more jagged, we might see more of this 'less intelligent overall but useful for the frontier tasks'-type models, meaning picking a model will only get more difficult from here.

    (FD: I work at AA)

    • GodelNumbering 7 hours ago

      > (FD: I work at AA)

      I do apologize in advance for putting you in spot like this, but there are two things I have been quite curious about.

      1. Since 60% of the benchmarks contributing to AA are now private only, how do you guys ensure transparency? More precisely, how can someone verify the transparency in scoring?

      2. There are also disclosures about custom benchmarking services offered to AI labs, what does AA do to keep the custom benchmarking services not getting access to the same/similar data that is used in the private benchmarks?

    • seizethecheese 7 hours ago

      Okay, this "better at long horizon tasks" thing might explain why Carbon scores worse on most benchmarks yet is the larger model.

      I looked into this using data I have from a battle system built on top of OpenCode (so we're trying to capture real world use). It looks like Carbon actually does a lot better when it uses significantly more output than opponents.

      Win rate when output is less than 2.5x opponents: 53% (41/78)

      Win rate when output is more than 2.5x opponents: 73% (41/56)

      The p-val on this is 0.02. Data from https://agentduel.com/leaderboard.

      There doesn't seem to be any relationship between actual size of task though. Win rate appears random based on total output of Carbon and also opponents. (Though our data is still small, so it might exist.)

      Looks basically like Carbon was trained to think a lot longer, not necessarily that it's better on large tasks.

    • abixb 8 hours ago

      Yes, I feel the next differentiation in model capabilities would be wrt how 'long of a horizon' goal they could autonomously work toward without breaking. We're in need for a new generation of evaluations and benchmarks for this -- METR doesn't cut it anymore.

    • seizethecheese 8 hours ago

      That would clear up why Carbon is such an outlier in terms of output tokens per task.

    • WarmWash 8 hours ago

      Neon is probably distilled from Google's internal successor to Carbon.

    • drivebyhooting 8 hours ago

      What we need is a model that automatically delegates intelligently (via RL and benchmaxxing?) so I can be freed of this burden.

      • baoooooooooooo 7 hours ago

        This comment always comes up but isn’t this literally irreducibly complex in the sense of Kolmogorov complexity?

        I’m out of my depth here but it does seem prompts have a “smell” to their complexity that means we can use a classifier to mate them to a model with better success than random noise, but still don’t essentially all of the savings associated with systems like this just come because there is some “complexity arbitrage” where there is enough of a token cost differential between models that you always save money by trying the cheap model first and then trying a hard model if it turns out to be hard?

      • nicoburns 7 hours ago

        This is being done at the layer above. "routing" is the term to google.

        • drivebyhooting 6 hours ago

          Routing at the prompt level seems naive. How is that better than a janky classifier.

          What I want is a deeply introspective model that knows to take a step back and call in for aid.

          • nicoburns 6 hours ago

            It's being done at the harness level. It's still a model making the call.

      • iamdelirium 8 hours ago

        Antigravity has a Orchestrator model that does this.

      • refulgentis 8 hours ago

        There's stuff that does this and I've put a substantial amount of time last couple months looking into it, unfortunately you always, always, end up at chicken-and-egg problems.

        It's really frustrating to describe. Easiest is you can't score on just the prompt, you need the full context. Ex. "Solve Fermat's last theorem" initial prompt then "continue" followup prompt. Which gets you back in the tarpit of "if I have to pay a model to do an inference run on context, then, the context is cached and I should continue using that model."

        The vision of another model thats RL maxx'd on knowing how to delegate has a few problems. One, we can observe externally, it died/didn't work even at OpenAI's scale and willingness to make it work, i.e. the...GPT-5.0? IIRC...router that silently went away. The other, if such a model was built, it'd also be good at regulating it's own thinking enough that it'd naturally use less tokens.

        In practice, the frontier hackers have settled on a decision model (i.e. Jev) x calibrating effort level or choosing from a set of models x reasoning effort tuples they specify rules for choosing in their Jev prompt. My guess is this stays stable and the only real choice.

      • canadiantim 2 hours ago

        Aren't people using jev or similar for decision-making like this?

    • chr15m 3 hours ago

      Does this mean models are not generalizing like the labs hoped?

    • mi_lk 8 hours ago

      what's AA

  • nicoburns 8 hours ago

    Is there a way to determine "model size class" from Google's naming scheme?

    • onlyrealcuzzo 8 hours ago

      I suspect they may want to get away from the size class due to the fact that - at the pace they are launching new models - and the similarity of model sizes between Neon/Carbon, Opus/Fable, Sol/Astra - the best model is going to be the most recently released one, not the biggest one.

      I'm not sure it's much better than 4.1 Flash vs 4 Pro -> if 4.1 Flash is better temporarily when Pro is typically significantly better, people are going to be confused.

      I'm not really sure how you get around this problem easily.

      It's not going to age well when the next Flash-lite model launches, but maybe they will call that `Gemini 4 <Element>-lite` - who knows...

      • nicoburns 8 hours ago

        Interesting. You're claiming that the Opus/Sol/Neon class and Fable/Astra/Carbon class models are similar sizes to each other? That would be surprising to me.

        I've found that while Opus 5.5 is often pretty nice for practical work, Fable (even 5.0) is still significantly smarter on hard tasks and makes far fewer "dumb mistakes"

        • dannyw 8 hours ago

          A Microsoft Azure page briefly claimed Sol runs with less looped transformer passes; and so you could speculate that Sol and Astra might share weights; but have different “loop” factors.

          That would make a lot of sense from an inference optimisation perspective. The same prefill and decode machines can hold one set of weights, and serve both Astra and Sol requests.

          At (multi-)terabyte-scale model weights, and when you have to offer not just the latest like Fable 5.2 but still serve 5.1, 5.0, etc; heck I think Opus 4.6 is still around; you really care about overall compute utilisation.

          Idle GPUs are wasted GPUs; and even over the fastest interlinks you’re not hotswapping multi terabytes of weights.

          With MoEs, there’s probably stuff you can play around with varying the number of active experts; or other clever optimisations.

          Maybe Sonnet 5.5 could be just Opus 5.5 with a different router and less active params? Who knows; I’m just theorycrafting now :)

          • onlyrealcuzzo 8 hours ago

            > A Microsoft Azure page briefly claimed Sol runs with less looped transformer passes;

            This doesn't really make sense given how the models released.

            OpenAI released Astra long before latent-space looping was a thing at the frontier space.

            Then they released Sol 6, which was fast, clearly not looped. Then they release Sol 6.1 which was nearly as good as Astra, but way cheaper, and WAY slower -> this would follow it just being a looped version of Sol 6.

            It doesn't seem likely that Astra 6.2 is just a looped Astra 6... It's too fast, and they'd go bankrupt if they're charging the same price... It may be a further looped version of Sol than Sol 6.2 - which would explain why it's not as much better as we would've hoped.

            It may also explain why their release is not going as well as they hoped. It's hard to predict how your first looped models are going to fare.

            If you look at how long it took Google to recover from their failed 3.5 Pro, it would be surprising if GPT recovered from their rumored failed run on Astra in under a month. That would give them reason to release a looped Sol and call it Astra.

            • iamdelirium 8 hours ago

              Looped transformers have been known for years now. I don't think it's something new.

              • onlyrealcuzzo 8 hours ago

                They are JUST arriving to frontier models.

                Even in the open weight Chinese models as small as DeepSeek V4.1 Flash - nothing is looped.

                • refulgentis 8 hours ago

                  I think they're responding to the claim Astra can't be looped, there's clear and convincing research, as well as public comms, that it is. It's unclear why Astra can't be looped because looping is new, but, 6.1 Sol released a week later can be.

            • refulgentis 8 hours ago

              Astra was the first shipping looper. Unconvincing evidence would be reporting x OpenAI's quick reaction to it via a blog post from their chief research (it broke what's considered a safety taboo). Convincing evidence would be research from ex. Neel Nanda on serial reasoning, recent update at https://x.com/EpochAIResearch/status/2122762411287748696?a=J...

          • nicoburns 8 hours ago

            I heard rumours that Sol 6.1 was cut down Astra because Sol 6.0 was received badly. But I thought that was the exception (and a result of Open AI's actual Sol-class model being uncompetitive) rather than the rule.

            • willy_k 8 hours ago

              That would line up - on release and for the few following weeks Sol 6.1 felt like Astra but maybe more task oriented, recently it’s been a noticeable downgrade in the level of specification needed to successfully set off.

        • onlyrealcuzzo 8 hours ago

          Fable & Opus, Astra & Sol, and Carbon are all thought to be >1T param models - AFAIK, most people think Opus & Sol are ~5T param models and Astra, Fable are 10T+.

          I think the rumor-mill consensus is that Google's models are significantly smaller while attempting to swing at the same level - but the ratios AFAIK are thought to be around the same - i.e. Neon is somewhere around half as big as Carbon, big enough that a newer checkpoint can make it temporarily a little bit better.

          • ml-anon 8 hours ago

            So many figures and ratios pulled from a single ass

        • weitendorf 7 hours ago

          There's only a limited amount of hardware for training and serving models at scale, and which form an upper bound for what model size/inference can support.

          A GB300 rack has 72x288GB=20TB VRAM. Do the math and 10T parameter models are just about the max size you can easily train and serve (at 1M context) without doing a lot of custom hardware development or limiting your scale.

        • andai 8 hours ago

          The Bitter Lesson - Richard Sutton (2019)

          https://news.ycombinator.com/item?id=19393432

    • XCSme 8 hours ago
      • jeffbee 8 hours ago

        You violated the only constraint the question contained.

        • ehsankia 8 hours ago

          Honestly, the bigger question is, are they at some point start adding numbers (Carbon 2, Neon 1.5, Hydrogen+ ionized), or are they keep giving a new name to every single new model. Argon seems to have been basically "replaced" by Carbon.

          • kzrdude 8 hours ago

            I hope they will continue to use numbers. Right now it's "Gemini 4 Hydrogen" in the UI, that seems to be the way it will be used.

          • lukax 8 hours ago

            Carbon is just Argon with extended thinking.

          • andai 8 hours ago

            Subscribing for Gemini 19 Bismuth

        • XCSme 8 hours ago

          Think of it like a riddle: given the order and naming scheme, OP has to imagine what Google thinks when naming the models...

      • onlyj 8 hours ago

        Why does this chart not include the other gemini 4 model, Argon? Also I am sure somebody thought about which element to name which model and it would be very helpful to me if they would explain their reasoning. OpenAI's naming got kind of abstract with 5.6 but I could follow it: Luna(Moon) was smallest, Terra (Earth) was medium, Sol (Sun) was biggest. Astra kind of breaks from that but whatever. But like I would genuinely appreciation someone from google to explain the reasoning behind their element names. Like are Argon and Neon related in some way because they are both noble gases? Is hydrogen the smallest model because it is the lightest element? Carbon has a lower atomic number than neon but is more expensive and presumably bigger... maybe I am overthinking this.

        • thevillagechief 8 hours ago

          Argon is already deprecated.

          • _delirium 7 hours ago

            Inconsistently it seems. On the $20/mo tier, Argon has been removed from the web interface and app, but is still available in Antigravity (while Carbon is only on the $100/mo tier in both).

          • LewisOrton 2 hours ago

            is it supposed to be replaced by neon?

          • jeffbee 6 hours ago

            Uh yeah, is this the big news of this announcement? They are moving all the good coding models into the Ultra price tier.

            • Ternari 5 hours ago

              Are they? Neon seems great and it’s available on the Pro tier.

              • jeffbee 5 hours ago

                I'm going by the model labels in Gemini.

      • troupo 8 hours ago

        Leave it to Google to create a card that goes from highest to lowest, and not the other way around

    • daemonologist 8 hours ago

      I have to assume they're messing with something other than size here - nobody would be this bad at naming different model sizes. Maybe sequence length or post-training strategy or sparseness or something else with the architecture.

      (Although why they'd release all ~four is a mystery. I guess they're having a hard time picking the "best" one and so are just putting them all out to let the public decide.)

    • ml-anon 8 hours ago

      “Model size class” is a completely meaningless taxonomy.

      It doesn’t correlate with intelligence, speed or price furthermore you have no way of telling what the model sizes of “like for like” named models from oai or Anthropic are.

      • NiloCK 3 hours ago

        Shouldn't it correlate pretty strongly with price? Like, model size is literally the amount of memory / compute it takes to elect each token.

        Agree there is no direct relation to quality - a Fable whose weights have been totally randomized would be the same size and cost the same amount to run.

      • sdenton4 8 hours ago

        /me pours one out for neural scaling laws

        • ml-anon 8 hours ago

          Did you read and somehow fail to understand the chinchilla paper?

          • sdenton4 5 hours ago

            Do you believe the amount of available training data has stopped growing?

    • DonsDiscountGas 8 hours ago

      I assume atomic weight. Neon seems to be better than Carbon, also faster and cheaper (all judging by the link)

      • daemonologist 8 hours ago

        What about Argon though?

        • breakwaterlabs 8 hours ago

          Apparently even Google wants to pretend it doesn't exist.

      • dudeinjapan 3 hours ago

        Neon is more noble.

      • Tadpole9181 8 hours ago

        But Argon has a higher atomic weight than both, and sits below Carbon and above Hydrogen. Also, Carbon is rated at better at long-horizon tasks while Neon seems to be smarter at contained work. So they're not strictly better/worse.

      • ehsankia 8 hours ago

        Great, consumers now need to first learn chemistry to be able to figure out which model to use!

        • accrual 8 hours ago

          That's my thought too. If one needs a lookup table to understand the product by name then it's a product issue, not a consumer issue, even if everyone "should" have the first N elements memorized by atomic $property.

    • moonu 8 hours ago

      melting point/boiling point

      • ahofmann 8 hours ago

        Ugh, this sound so overly nerdy, that it could be very true. I do not follow or care about the Google LLM releases just because the naming scheme is so convoluted (and the constant churn of products around the llms). I learned the anthropic naming scheme, it is stable and easy to follow. And this is why Claude is my daily driver. I don't like change, I like stable and predictable progress. Open AI naming scheme is kind of easy to learn (bigger planet = bigger model). But fuck Googles bullshit naming crap, seriously.

    • whimsicalism 8 hours ago

      meant to be an opus/sol competitor, looks like. not fable/astra

    • catigula 8 hours ago

      If there is, it isn’t obvious, which makes it terrible.

    • drum55 8 hours ago

      The developer guide lists

          Gemini 4 Neon (NEW!)
          Gemini 4 Hydrogen (NEW!)
          Gemini 4 Carbon (NEW!)
          Gemini 4 Argon (NEW!
          Gemini 3.8 Live 
          Gemini 3.8 Flash (NEW!)
          Gemini 3.7 Flash
          Gemini 3.6 Flash
          Gemini 3.5 Flash
      
      I've no clue whatsoever how to read anything about the names of them and what they might imply you would use them for though. Within these most of them have LOW MEDIUM HIGH XHIGH, with the default being different for some of them, how anybody is possibly meant to understand this whole matrix of ambiguously named products is beyond me.
      • elxr 2 hours ago

        Don't use Argon or Carbon.

        Hydrogen is the cheap one for straightforward tasks.

        Neon is supposedly Opus-level, which is the exciting one if it actually works as well as they claim.

        Obviously anything with a 3.x is not what you should reach for. This has always been the case on a first digit bump in any of these mainstream models.

      • aenis 8 hours ago

        Same here. First i thought it would be noble gasses. Then I thought it would be about the atomic weights. I know, I should just ask AI to figure it out for me, but its a bit humiliating.

        • drum55 8 hours ago

              Hydrogen 1
              Carbon 6 
              Neon 10
              Argon 18
          
          Maybe this implies that Neon is smaller than Argon in weight size, but it's newer, but it's also only a little bigger than Carbon? Do the Nobel gas models have more safeguards? Is Carbon 6x the size of Hydrogen and a third the size of Argon? When they release Xeon can we make compounds? Feels like some really cheap Da Vinci Code esque thing if we're supposed to get some sort of meaning out of the names.
          • svnt 8 hours ago

            From their Neon marketing:

            > In our early evaluations, it matched or exceeded Gemini 4 Carbon on overall quality at 3x the speed and 60% of the cost. We’re excited to bring these capabilities to Glean customers tackling multi-step work that demands accuracy, detail, and speed.

            So possibly smaller than Carbon? Or some other axis that is making it faster and cheaper? No idea.

            • aenis 8 hours ago

              Same. Running my benchmarks to verify their quality claims. Carbon performed very well on my private suite - spitting distance to Opus 5.5.

              • aenis 3 hours ago

                ...first runs are in, and Neon is far below Carbon in terms of quality, but indeed much faster - ~4x faster, and 2x cheaper per task. Probably a good model for delegating to after key design work is done with something cleverer.

                • svnt an hour ago

                  What category of work did it underperform on for you? I am seeing comparable outcomes in some technical document processing, but the cost is not lower for me.

        • iainmerrick 8 hours ago

          It could be worse, they could have named them after different Google chat products.

        • jimbokun 8 hours ago

          How do you decide which model is best for telling you which model to use?

      • prymitive 8 hours ago

        It’s just a matter of time before we have

        - Gemini (New)

        - Gemini

        - Gemini (original)

        • murderfs 8 hours ago

          This is Google, it'd be:

          - Gemini (Not ready yet)

          - Gemini (Deprecated)

        • accrual 8 hours ago

          They could take a page from Nintendo: New Super Gemini Deluxe!

        • abixb 8 hours ago

          Inb4 they switch it back to Bard.

        • jimbokun 8 hours ago

          Classic Gemini

      • kzrdude 8 hours ago

        Hydrogen is the new flash, at least

    • Mogzol 8 hours ago

      As far as I can tell it's arbitrary and Google is just choosing element names they think sound good.

      When Argon first came out it seemed like they could be going for noble gasses, like Helium (small), Neon (mid), Argon (big).

      But then Carbon came out and it seemed like maybe they were going through the alphabet for model generations. Argon, Boron (internal), Carbon.

      And now they've come out with Hydrogen and Neon, and there's no logical way to order them. The actual model sizes are Hydrogen (small), Neon (mid), Carbon (big). This doesn't match sorting by atomic number or alphabetically.

    • kphorn 8 hours ago

      They use the very precise "it sounds cool". Neon is 75% the price of Carbon but has 4 extra protons in the nucleus and is also a noble gas. Argon is a heavier noble gas at 16 protons.

      element/model - atomic number - cost H - 1 - $1.50 & 7.50 per mtok C - 6 - $4 & 20 per mtok Ne - 10 - $3 & 15 per mtok Ar - 16 - $2 & 10 per mtok

      Claude - length of story but fable messes it up Haiku - 3 lines Sonnet - 14 lines Opus - a novel? Fable - 30 lines?? Mythos 15700 lines?

      Luna - 28 days Sol - 365 days Astra - 5-6000 stars

    • WarmWash 8 hours ago

      They are almost certainly being ambiguous so that people rely on the model descriptors to pick which one to use.

      I think there are a ton of people who automatically default to the largest model with the highest thinking for everything. Which historically made sense but not nearly as much anymore.

      At this point a model like hydrogen can cover just about everything an average person is doing. You don't need Carbon deploying an agent swarm pushing terabytes of parameters through HBM memory to help you plan your birthday party.

      • jeffbee 8 hours ago

        Wouldn't such users immediately run into quota walls?

        • WarmWash 8 hours ago

          They do, and then they complain.

          Worse, people burning dumb prompts on large models also has the net effect of shrinking quotas for everyone.

    • adithyareddy 8 hours ago

      (S) Gemini 4 Hydrogen - fastest - $1.50/$7.50

      (M) Gemini 4 Neon - second fastest - $3/$15

      (L) Gemini 4 Argon - second slowest - $2/$10 (till Nov 30), $4/$20 (> Nov 30)

      (XL) Gemini 4 Carbon - slowest - $2/10 (till Nov 30), $4/$20 (> Nov 30)

      • WarmWash 8 hours ago

        Argon and Carbon are the same class model. Carbon is just a newer rev of it.

        There is a Helix model too (in the pipeline) which is the actual small size.

    • baoooooooooooo 8 hours ago

      Everyone seems quite upset/perplexed by the Gemini naming scheme but I quite like it. And hey, if you find yourself lost, the periodic table is probably the most ubiquitous ordered list in the world and serves to clear things up pretty fast. This will likely not be strictly true as models release over time, but is a good rough indication of what the team thinks we should expect of the models, in the same way the other labs run their naming.

      I also like the personalities implied by these, with Anthropic saying “human literary achievement is neat”, OpenAI saying “space is neat” and Google being like “science is neat”.

      Edit: I was coming back to say “I wish we could talk about stuff more substantive than the naming scheme” but (1) my comment may be objectively incorrect, and (2) I have to apologize that I haven’t helped the situation by… commenting on the naming scheme.

      • WarmWash 8 hours ago

        They are not following the periodic table in any discernible way though.

        • dudeinhawaii 7 hours ago

          They're in alphabetical order, almost like Ubuntu release names. There was a leak about a Gemini Oxygen model, which would further support alphabetical order -- skipping names they don't want to use.

          So, names don't align with capability but instead "release freshness". This is unlike every other provider but it is what Google does (did?) with Android releases I believe (dessert names).

          • WarmWash 6 hours ago

            They went form A to N in 4 releases. Initially maybe they planned to follow the alphabet, but I think that went out the window

      • georgemcbay 8 hours ago

        > And hey, if you find yourself lost, the periodic table is probably the most ubiquitous ordered list in the world and serves to clear things up pretty fast.

        But the way google ranks the models by capability doesn't follow periodic table order.

        Hydrogen - 1

        Carbon - 6

        Neon - 10

        Argon - 18

        It could very well be based on their density (g/cm^3 in standard conditions):

        Hydrogen - 0.0000899

        Neon - 0.000900

        Argon - 0.001784

        Carbon - 2.267

        This would make more sense than assuming it is based on their periodic table order as it more accurately maps to how Google markets their capabilities, but is too esoteric to really be useful for clearing up confusion for normal users, IMO.

      • wat10000 8 hours ago

        The periodic table isn't really helping. Hydrogen is the lightest, that much makes sense. Beyond that it doesn't seem to work. The elemental ordering is H C Ne Ar. The cost ordering is H Ne Ar C (with Ar and C being the same per token but I gather that C thinks more and thus costs more). The capability ordering seems to be H Ar C Ne. Argon is the highest on the table but is neither the most expensive nor the smartest.

      • pavo-etc 4 hours ago

        Lmao no. The most ubiquitous ordered list is... numbers. Or the alphabet.

    • genxy 8 hours ago

      I think you mean say nobel size, I'd wait for radon.

      • swasheck 8 hours ago

        now we have to know our atomic weights to be able to choose which gemini model for the task, i suppose.

      • CamperBob2 8 hours ago

        Problem with Radon is that they will swap it out for Polonium without telling you, and then they will obsolete that in favor of Lead.

  • seizethecheese 8 hours ago

    It looks like it's being served extremely fast via Openrouter right now https://openrouter.ai/google/gemini-4-neon. 127 tokens per second. This is way faster than Carbon at 46 tokens per second (https://openrouter.ai/google/gemini-4-carbon). I'm a bit confused by this, since it seems like Neon is supposed to be their flagship model. Perhaps they've been throttling Carbon?

    Carbon has a huge problem that it's just way too slow, and outputs way too much per task (on real world tasks, it's the worst at this https://micjm.com/blog/carbon-is-that-slow/).

    Without the speed issue, Carbon would be a top model already. It's top on our battle system built on opencode, if you exclude early votes: https://agentduel.com/leaderboard?excludeEarlyVotes=true, though to be fair, small sample size so far.

    Conclusion is that it's looking good for Neon. Even if it suffers from similar problems as Carbon, if it's even smarter and way faster, it should be a top model, maybe the best.

    Edit: I was wrong about Neon being the larger model. I made this assumption because Neon has higher scores and they're both "v4".

    • WarmWash 8 hours ago

      Neon is an Opus sized model

      Carbon is a Fable sized model

      Carbon seems to be a version of Argon that is incredibly spendy with thinking tokens.

      • seizethecheese 8 hours ago

        Okay, that clears it up. Their naming convention is confusing because Neon has a larger atomic number than Carbon, so there's just no pattern here at all.

        Weird because Neon is outscoring Carbon in AA index, Terminal bench science, etc and they're both named "4". Should have released this as Neon 4.1.

        With that clarification, I'm sure Neon has the same overthinking problem.

        • WarmWash 8 hours ago

          Neon is weeks newer than Carbon, hence the improvements.

          Google also has a rumored Krypton internal model, which is also Argon/Carbon class. Neon might very well be distilled from Krypton, not Carbon.

          • seizethecheese 8 hours ago

            It's not weeks, Carbon was released on the 10th, 8 days ago.

            > Neon might very well be distilled from Krypton, not Carbon.

            That would make a lot of sense, interesting!

            • WarmWash 8 hours ago

              Right, but Google has had Carbon internally since early October IIRC.

          • desterothx 7 hours ago

            in the model card they pretty much say it's based on carbon

        • elxr 2 hours ago

          AA index doesn't tell you much. Some small models with low AA index are still incredibly useful because you know they can stay coherent and behave well over long tasks, and use your harness effectively despite the lower supposed "intelligence".

          Some giant models overengineer things and add redundant checks everywhere instead of simplifying overall design. In humans, that'd be a sign of low intelligence.

      • aenis 8 hours ago

        [citation would be appreciated]

        • WarmWash 8 hours ago

          Google (and the other labs) don't give raw model sizes. It's inferred from price/speed/benchmark/quota. When you align those parameters, it's clear what models are competing with what.

          Also neon seems to be stepping on the toes of Carbon a bit, just like Opus 5.5 did to Fable.

          • seizethecheese 8 hours ago

            Well, Anthropic named Opus 5.5 and Fable 5.1 then 5.2 to indicate that the smaller model was flying ahead.

            • WarmWash 8 hours ago

              I'm almost positive google is purposely trying to breakaway from names describing sizes. If it works (read: helps balance demand), then the other labs will probably do it too.

              • syradar 3 minutes ago

                Well it’s making me not wanting to try their models even more.

              • desterothx 7 hours ago

                it would be funny if they did some marketing around the hydrogen model (like as powerful as the h bomb) to try to push more non coders to use it for normal tasks

    • Melatonic 8 hours ago

      I wonder if this is an experiment with their new partial chip design that's vaguely similar to Taalas

      • seizethecheese 8 hours ago

        It looks like it's actually a smaller model than Carbon, just really confusing release.

    • elxr 2 hours ago

      > Carbon has a huge problem that it's just way too slow

      And makes shit up, and costs significantly more than Sol/Opus on tasks requiring lots of decision making.

  • onlyrealcuzzo 8 hours ago

    This model feels like it might actually be as good as Opus 5.5.

    This is probably the set of benchmarks most people care about: https://www.vals.ai/comparison?m=claude-fable-5-2,cobalt-c-h...

    Overall, performs just under Fable (above Astra or Sol 6.2 or Carbon - slightly ahead of Opus 5.5).

    Latency is by far the best for the level of quality.

    It's about 2x more expensive than Sol 6.2, but significantly better and faster. You get what you pay for.

    • SuperV1234 8 hours ago

      I am a Claude Ultra subscriber that's very happy with Opus 5.5.

      If I wanted to realistically try out Neon in some of my workflows without committing to a paid subscription before seeing if it can actually work as well as Opus, do I have any option, or do I just need to take the gamble?

      • user43928 8 hours ago

        I was thinking of giving Gemini 4 Neon a try as a code reviewer instead of GPT-6.2 Sol, and found this in their FAQ:

        "Using third-party software, tools, or services to access Antigravity is a violation of our Terms of Service"

        and also: https://antigravity.google/docs/faq?can-i-use-antigravity-cl...

        I'd like to keep my Gmail account, so Google can keep their model and will not get my business.

        • WarmWash 6 hours ago

          You can't use subsidized tokens in 3rd party apps. Although they just suspend your Gemini account now, not your google account.

          That being said, you can do whatever you want with API tokens.

          • user43928 5 hours ago

            You can with Anthropic and OpenAI, using 'claude -p' or 'codex exec'.

            Paying for Gemini at API rate is therefore out of the question for me.

        • fastest963 6 hours ago

          I think that page is out of date. You can use it via T3 Code now and Google has said it's okay. It's via their official ACP-compatible endpoint that's now provided by the antigravity-cli.

        • koolala 3 hours ago

          You can install Codex in Antigravity its kinda fun running both.

        • what 7 hours ago

          You don’t need to use the models through antigravity. You can get them through ai studio (or whatever Google calls it) and pay api pricing to use it via whatever harness you want.

      • layoric an hour ago

        How do you find Fable 5.2? I tried it for a few days for coding, and I'm back on Opus 5.5, and even Deepseek 4.1 Flash for simple stuff. It was just going off on tangents way more for my workflow.

      • onlyrealcuzzo 8 hours ago

        They say it's going to be available for Pro users - so you don't need a $200 account (Ultra). But, I don't think you're going to be able to try it unless you get at least a $20 account.

        IIUC, you also don't get to try Opus unless you have a subscription, right?

        • deaux 5 hours ago

          Reminder that Google trains on your data if you're using a consumer sub whereas Anthropic and OpenAI claim they don't if you've got all relevant boxes turned off. I know everyone stopped caring about this at the start of the year and believes the latter two don't honor it anyway, but it's good to know.

      • what 7 hours ago

        You can always pay API pricing if you don’t want a subscription?

      • desterothx 7 hours ago

        openrouter or api directly? great question to use your claude sub to figure out easily :)

    • catigula 8 hours ago

      0% chance this is better than any frontier OpenAI model.

      • WarmWash 6 hours ago

        I think that OAI likely has the strongest internal model, but I don't think they can figure out how to ensure alignment on it

        • f13f1f1f1 5 hours ago

          I also have the hottest wife but no one can see her because people wouldn't be able to constrain their lust. I need to hide her to maintain safety.

          • WarmWash 5 hours ago

            However hot she is, she isn't solving millennium problems.

      • HDThoreaun 8 hours ago

        How can you be so confident about a model that was released today? I just dont get the lack of epistemic humility here

        • catigula 8 hours ago

          >how can one presume the sun will come up tomorrow merely because it has come up every day prior? have some epistemic humility! i am a very smart rationalist.

          Edit: I tested the model and it found several mistakes Astra 6.2 made. It’s possible…

  • MisterBiggs 5 hours ago

    I've been incredibly impressed with this model so far. Argon felt like a leap in presentation, the model writes really good documents and replies. Neon feels like another big leap in usability over that, the model seems pretty smart but its super eager and runs down tasks really well and efficiently. Definitely passes the vibe test

  • ByteAtATime 8 hours ago

    This is an interesting model - I feel like they're trying to avoid OpenAI/Anthropic's approach of various model tiers, and just keep climbing the Pareto graph (for lack of a better term) as fast as possible, both for cost and performance. I think I personally like the approach, despite the confusion it causes.

  • bemtevi 8 hours ago

    If one day, AI becomes provably good at all programming tasks, at working at mathematics better than all mathematicians, and at solving problems in almost all domains with maximal efficiency, I think it'd be a dystopia.

    What I simply don't get is why people want to push the boundaries of efficiency and knowledge. I used to think as a teenager that all knowledge was good but how good is knowledge when we can't acquire wisdom with a comparable speed?

    The greatest thing about the early internet and earlier science was that people shared and communicated with each other. Now they just share what they produce wihouth much humanity at all.

    It seems to me that the creation of all these models has just become a race where no one really cares about the finish line and no one really understands why we're doing it, except for the short-term trail of dollar bills on the race course.

    For me, each new model release is a sad day for humanity.

    • simpsond 2 hours ago

      We are already there, and yes, it often feels inhumane.

  • dannyw 8 hours ago

    Major props to Google for announcing a model and releasing it on the same day.

    Unfortunately, their recent change to allow me to opt out of AI training (which would also include connector data like Gmail, photos, etc) hasn’t rolled out to me yet[1]; so I won’t be upgrading my subscription; nor using the Gemini app until I can opt out.

    [1]: https://arstechnica.com/ai/2026/11/google-will-let-you-opt-o...

    • sdvs9so 7 hours ago

      I think that just changed as of a Nov 16th. You're now able to opt-out your chats from model training while retaining chat history[1].

      [1]: https://blog.google/innovation-and-ai/products/gemini-app/up...

    • 5555watch 8 hours ago

      "Gemini History" seems to only target Gemini App functionality, what about Antigravity chats?

    • deaux 5 hours ago

      > Currently, if you disable history, Google still retains that data for 72 hours (and 24 hours for the user). Once the updated page is live, this duration will change to two months.

      So a regression then.

    • testtestmetest 8 hours ago

      Oh know, they're going to steal the code for you 3 MAU SaaS !!!

      /s not trying to be mean, its just weird hill to die on. Every company is stealing all your data in ways you unknowingly consent to at every second of the day.

  • stillpointlab 7 hours ago

    I'm getting fatigued with the constant introduction of names and the impenetrability of mapping these brand new names to use cases.

    I think internally, Google, Anthropic and OpenAI are thinking "It's so easy, you just use Neon for X, Carbon for Y" - but most of the world is trying to map Sol, Astra, Neon, Opus, Fable, Mythos, Carbon, Haiku, etc.

    I just hope they finally land on a naming scheme they are happy with and just stick with it. But I have my doubts. Perhaps they are all benefiting from this confusion.

    • aboardRat4 7 hours ago

      >m getting fatigued with the constant introduction of names and the impenetrability of mapping these brand new names to use cases.

      Because probably there is no system. They just train a model, and find out that it's "kinda good", and that's it, they publish it. The models do not constitute even a partially ordered set, nobody can predict their ranking for your particular tasks.

      • motbus3 6 hours ago

        Like some vendors, making the name scheme incomprehensible is part of you not knowing which one is the best cost benefit

        • captainbland 6 hours ago

          If you have to experiment to find out, you have to spend tokens

        • elxr 3 hours ago

          Test it then.

          Comprehend it using real use, or find someone you trust to use them in anger, don't wait around for the labs to do everything for you. You're the one with the leverage on where you send your money to.

          • matkoniecz 41 minutes ago

            Maybe if you spend millions per year on tokens it makes sense to test all of them.

            But at rate how many models appear it would cost more for me than potential savings.

          • jychang 2 hours ago

            Nah, that's like trying to test drive every single car on the market, in order to buy a car.

            Just buy from companies with non shitty naming systems like Anthropic, and let the shitty names go bankrupt. Free market and all that.

        • paradox460 6 hours ago

          Wonder when Sony is going to release a model, and pull it's name out of a Scrabble word sack

          • mrandish 5 hours ago

            True story about Sony naming: I was at the PC industry trade show where Sony launched the new VAIO computer in a big, beautiful booth at the front of the hall. Being a long-time Sony fan, near the end of the first day after the crowds had mostly cleared, I stopped by the Sony booth on my way out to see the new PC. I saw that the VAIO name also formed the computer's distinctive ultra-modern logo featured in large bold text on each side of the PC's futuristic case in a minimalist block font. But I was puzzled how a company with a history of so many great products with awesome names (WalkMan, DiscMan, PlayStation) had come up with such an awful name.

            Since I was curious about the origin of the name, I wandered over to where three Sony employees were chatting passing time until the day ended. I learned all three were mid-level marketing staff based in Sony's New Jersey US HQ. So I asked "How do you pronounce the computer's name?" All three had different takes between various long vowel vs short vowel sounds.

            Since they were all facing me, I also asked, "Without turning around, how do you spell the name?" They all paused, then laughed. They did manage to work that out after a few false starts but it wasn't exactly smooth. As we chatted some more I asked who'd come up with such a unique name and learned Sony had hired a "really cool" high-end branding agency who specializes in coming up with great product names. The effort and expense in creating the name and logo had clearly been marketed to employees as "cutting-edge cool." I warmly wished them the best with the new product before wandering out of the hall to dinner.

            • teshigahara 4 hours ago

              None of it true of course. The VAIO name and logo was by Gotō Teiyū who had been working as a designer at Sony for 30 years at that point. Its pronounciation is quite unequivocal in Japanese as well (like bio).

              • mrandish 4 hours ago

                Well, I guess the mid-level marketing people in Sony's New Jersey office got some garbled version of what happened in Japan, because that's what they told me at PC Expo '96 at the Javitz Center.

                Regardless, the PC and it's neat industrial design plus Sony's brand and distribution were good enough to overcome a weak name. VAIO was an acronym for "Video Audio Input Output" (which Sony later changed to "Visual Audio Intelligence Organizer", although I'm not sure why they thought that would be better).

                • teshigahara 4 hours ago

                  Sure, I meant none of what they told you was true. I typically wouldn't expect much better from sales people.

              • lejalv 17 minutes ago

                Yes, but, you know, the US is the center of the world

            • grantsucceeded 20 minutes ago

              i have to say, that ultraportable laptop was mind blowing to me at the time, maybe still.

      • floatingtorch 4 hours ago

        There is definitely a system. They are training models of different sizes for different levels of inference efficiency, which maps to token pricing. The ordered set is the size of the weights/ computational complexity during inference.

      • ijidak 2 hours ago

        I feel Anthropic and OpenAI are actually doing a great job with their naming.

        We intuitively understand that Luna is smaller than Terra, and Terra is smaller than Sol, and Sol is smaller than Astra.

        Likewise, we intuitively understand that a haiku is shorter than a sonnet, and a sonnet is shorter than an opus, and an opus is shorter than a fable. Fable is probably the weakest of their names. Maybe they should have gone with Epic.

        It's Google that isn't doing it right.

        Even those who know physics have zero intuition regarding carbon, argon, and neon.

        Hydrogen being the smallest makes sense.

        But what property is the point of comparison for carbon, argon, and neon?

        Valence shell? Atomic mass? I saw someone throw out density.

        It provides no self-evident point of comparison.

        Frankly, if they can't even get naming right, they're likely in trouble on the productization side.

        It reminds me of Microsoft in the Zune era.

        They just don't have what it takes to win on the product side in this battle.

        • gfody 19 minutes ago

          Sol should be 400x larger than Luna

        • wpm 37 minutes ago

          Intuitive would be naming them "small medium large" or "good, better, best", not depending on the customer to know Latin or the periodic table or fucking whatever.

          Hell, steal the terrible muddy naming from Apple and it would be better. Claude, Claude Pro, Claude Max. GPT-5.6, GPT-Pro-5.6, GPT-Max-5.6. Whatever! Tiering is good and fine, just make it clear.

        • Daneel_ an hour ago

          Huh, I guess there is an order. Definitely not 'intuitive' unless you actively try to reason about the names. I'd always assumed they were basically code names that had no bearing on size or performance.

          I only understand the Luna/Terra/Sol/Astra order though. I couldn't have told you that a sonnet or an opus or a fable had any sort of sizing guidance. English go brrr.

        • grantsucceeded 9 minutes ago

          sorry, but it might as well be small, medium, large. oh, and maybe supersize.

          its all so one dimensional, just call it moon, earth, sun and stars cause that will make us all warm and fuzzy.

          oh, no wait, it'll be way more informative and clear to call it terra,luna,sol and astra. Yeah, thats way more genuine caus it makes me feel smart for figuring out it means small, medium, large and supersize.

      • ijidak 2 hours ago

        I feel Anthropic and OpenAI are actually doing a great job with their naming.

        We intuitively understand that Luna is smaller than Terra, and Terra is smaller than Sol, and Sol is smaller than Astra.

        Likewise, we intuitively understand that a haiku is shorter than a sonnet, and a sonnet is shorter than an opus, and an opus is shorter than a fable. Fable is probably the weakest of their names. Maybe they should have gone with Epic.

        It's Google that isn't doing it right.

        Even those who know physics have zero intuition regarding carbon, argon, and neon.

        Hydrogen being the smallest makes sense.

        But what property is the point of comparison for carbon, argon, and neon?

        Valence shell? Atomic mass? I saw someone throw out density.

        It provides no self-evident point of comparison.

        Frankly, if they can't even get naming right, they're likely in trouble on the productization side.

        Their fumbles remind me of Microsoft in the Zune era.

        They just don't have what it takes to take first or second place in this revolution.

        But they'll likely do fine indirectly via TPUs, the Google homepage, Gmail, etc. Just as Microsoft was okay because of their legacy dominance.

    • josephg 4 hours ago

      With anthropic at least they have a naming system based on the size of the model, with literature analogies:

      Haiku < Sonnet < Opus < Mythos/fable

      I think openai has a similar system with astral bodies:

      Luna < Terra < Sol < Astra

      • mlinsey 4 hours ago

        I like both naming schemes...but the thing that always confuses me is, aren't fables pretty short? The Tortoise and the Hare is a lot shorter than say, Anna Karenina or other opuses...

        Although, maybe it's called Fable because it's trying to bluntly convey strong moral lessons, hence the strong safety classifiers...

        • nimbleal an hour ago

          I find Open AIs system confusing also because my mental metaphor for their names is a pre Copernican cosmos so it should go terra > luna not the other way around

        • amirhirsch 2 hours ago

          One thing is for certain: cache invalidation is decidedly the second hardest problem in computer science

        • mattnewton 3 hours ago

          It’s Fable because they used up the name Mythos internally already for the model class.

          • KingMob 2 hours ago

            They could have used Epic, but that's a heavily-overloaded term in software. Maybe Saga?

            • fastball 2 hours ago

              Saga would've been good. Or they should've just released neutered Mythos to the public as Mythos instead of Fable, and then renamed Mythos to be like mythos-unlobotomized or whatever.

            • cozzyd 2 hours ago

              Tome?

      • emaro 4 hours ago

        I like them too but it took a while for me to land, tbh. Mistral does the best job IMO: Mistral Small/Medium/Large. Slap a number on for the generation.

      • rafaelc 4 hours ago

        This doesn’t land for the vast majority of people. Maybe 1 < 2 < 3 < 4 would work better.

        • mlinsey 4 hours ago

          It makes sense if you understand that the size of a model is something important and meaningfully different than just how advanced the model is.

          You could argue that's a detail the user shouldn't have to concern themselves with, but I think it's appropriate to expect them to learn that and pick names which put that concept front and center. The various "model router" approaches for consumer tools like ChatGPT have been finicky and inappropriate for professional tools we spend a lot of money on.

        • Aeolun 4 hours ago

          I’m not sure the vast majority of their audience doesn’t get it. I feel like it’s arguably clearer than 1-4 would be.

          • Daneel_ an hour ago

            I strongly disagree. The order of numbers are universally understood since they're taught since childhood. The arbitrary order of literature terms and old-timey astrological names are not.

            Better still is simply small/medium/large/extra large, or low/medium/high/maximum.

            • tdhz77 an hour ago

              Or perhaps Size, version and year

              Claude L126 Claude L226 Claude L327

              Now we know which model is the latest and which version we are on in that year. Now give it a cute name for your blog. And when people use Intrepid, we will all be confused all over again.

              Claude S126-Intrepid Claude S226-Jaunty Claude S327-Maverick

        • KylerAce 4 hours ago

          But the number is for the model iteration not strength within a generation.

        • utopcell 4 hours ago

          So Opus 5.5 would be "3 5.5"? I don't think that's better.

          • mrheosuper 4 hours ago

            How about "Claude 5 0505" ? First number is tier(5 = opus, 1 is haiku, 3 is sonnet), next 4 numbers are version(2 for major, 2 for minor)

        • npunt 4 hours ago

          "What's the #1 Claude model?"

          "Number four, obviously!"

          • bloppe 4 hours ago

            duh, it's number Astra

        • elxr 3 hours ago

          Maybe if you don't know a damn thing about branding. Imagine saying GPT-6.2-3

          ???

      • snthpy 2 hours ago

        Exactly, so even if you don't remember you can deduce the order. I just looked up the atomic weights of Neon and Argon and apart from Hydrogen the Google models are all in the wrong order. Unless I'm supposed to be using density?

        Edit: Density seems to track, H < Ne < Ar < C

        • snthpy 2 hours ago

          Still seems the wrong way around from an antigravity perspective. Hydrogen should be the strongest and elevate you the most.

      • wejick 2 hours ago

        Anthropic doing a good job sticking on the same naming from the beginning.

        OpenAI's is still confusing for me tbh, but Google's is on different level!

      • okrad 3 hours ago

        Google was close with atomic numbers of elements but Carbon bucks the trend.

        Hydrogen 1 < Neon 10 < Argon 18 < Carbon 6

    • Jayakumark 3 hours ago

      They should standardize on naming like Medals for all models.

      Medal - Anthropic Claude, OpenAI, Google Gemini , Meta Muse

      Bronze - Haiku , Luna , Hydrogen, Ember

      Silver - Sonnet , Terra, Argon (Now defunct) , Spark

      Gold - Opus , Sol, Neon, Wonder

      Platinum - Mythos/Fable, Astra , Carbon, NA

      Diamond - ?

      One flaw is people using it will always try to stay on gold or above - in their mind, gold might be better so they may spend more when it’s not needed.

      Silver is all you need now for most of coding work.

      Edit: updated argon and neon categories on table above as Neon replaced Argon

      • danpalmer 3 hours ago

        The Gemini models here are not quite in the right places. Hydrogen benchmarks at a Silver, Neon benchmarks at a Gold/Platinum depending on what you're doing. Argon benchmarks at Silver/Gold depending on what you're doing. Neon is very much above Argon.

        I'm not at all familiar with the Meta models, are they correct, or just put in place based on market positioning?

      • Ethan_Mick 3 hours ago

        This was actually very helpful, thank you! I need this. These tiers tell me exactly how to compare models and where to use them.

      • davidd_1004 2 hours ago

        Carbon's price and performance are much closer to Opus than Fable. And Neon is really supposed to supersede Argon (in that same tier). That's what makes it so confusing.

    • mstibbard 7 hours ago

      I promise you the confusion is felt inside these companies too.

      • jtfrench 5 hours ago

        This sounds highly likely. Imagine trying to name and market a bag of weights, and arguing with a straight face why it’s better than another bag of weights.

    • mrandish 5 hours ago

      After looking over the landing page, my first thought was also about whether I wanted to invest the time to figure out where this new model fit in the relative hierarchies of comparable product stacks. Ultimately I decided "Nah", and aborted to the comments.

      • oniony 4 hours ago

        They have some interesting graphs linked. From these I learnt I probably should be using GPT Sol for its bang to buck.

    • pavo-etc 6 hours ago

      I maintain a table for all labs. each row is S / M / L / XL (maybe 2XL 3XL etc in future) and each column is the lab. The only way i can contextualise any of this.

    • overleaf 4 hours ago

      At work at Giant Tech Company, someone has started publishing a "WTF" guide every week that includes the decoder ring for model names vs. their capabilities. I just have my agent read it every week and update all my skills and rules appropriately to figure out which model to use for what.

      • ehnto an hour ago

        Wouldn't that mean your skills and rule outcomes differ every week?

    • SamuelAdams 5 hours ago

      They really need an industry term for each model. Think like mobile, tablet, laptop, desktop. Apple, Lenovo, HP, Google, and others all have their own versions of these, but in general the public has some idea of common use cases that each product will satisfy without delving into the details very much.

      • 0x69420 5 hours ago

        to be fair we have certain semi-common names like "flash" and "small", even though those tend to be relative to the size of their flagship model, which makes the terms somewhat meaningless out of context... but at the very least those do help when you're familiar with the given company.

    • crystaln 5 hours ago

      Yeah when I talk to people who don’t follow ai closely they are very confused, particularly about Claude where you have to pick a model by an arbitrary name. At least ChatGPT has a slider for model strength

      • f13f1f1f1 5 hours ago

        Anthropic has at least been consistent these other ones change it every fucking week. I honestly just think of models in terms of haiku/sonnet/opus level now because the other ones are completely inconsistent.

        • piskov 5 hours ago

          The issue with anthropic is a fable is smaller than an opus

          So they fucked up naming as well in this regard.

          • tintor 5 hours ago

            And you believe opus model has less parameters than fable?

            • vbezhenar 3 hours ago

              There's no "Fable model". There's Mythos. And it's certainly larger than Opus.

              Fable is Mythos with safeguards, basically.

              In fact, Mythos also includes safeguards, so at this point it's not really clear what's the point. You can apply for security program and there are three levels of access to the Mythos. I guess now it makes sense to just drop "Fable" name and go with Mythos which will just be implicitly restricted according to your access level.

              Or maybe they'll realize that there are abliterated openweights models, and restricting Anthropic models does more harm than good. I literally have to use other service for security audit of my software, because Fable rejected it.

            • slowin 4 hours ago

              No, they mean the models are named after a form of writing where the longer form = a bigger model. Except an opus is longer than a fable.

            • apetresc 4 hours ago

              Yes, almost certainly.

          • MikeTheGreat 5 hours ago

            To be fair, I think they wanted to name it "Mythos", not "Fable".

            Then the Trump administration panic-banned Mythos and Anthropic smartly decided to maybe call it something else.

            • trollbridge 2 hours ago

              Well, they ran around boasting about how dangerous Mythos was. Maybe they should call it Ragnarok

    • dimavolo 2 hours ago

      I’d like to see them use Apple’s naming scheme:

      Gemini Mini, Gemini, Gemini Pro

      GPT Mini, GPT, GPT Pro, etc

      And then just increase the version

      GPT Mini 6.2 = Luna 6.2

      Gemini Mini 5.1 = Gemini Neon

      Etc

      Maybe add a 4th/5th one in there sometimes, eg to distinguish Gemini hydrogen from Gemini Neon, like “Mini” vs “Air” does

      • jezzamon 2 hours ago

        They literally had this (flash, pro, etc) and then decided to discard it for some reason...

        At Google we make memes about how the naming system is incomprehensible

    • pizzafeelsright 3 hours ago

      I deal with this at a global level.

      There isn't yet a single good method. We're getting close to unifying the naming practice.

      The routing and harnesses are still not adapting fast enough but when they catch up the names won't matter.

    • minimumvector 3 hours ago

      What I can't understand is how analogous in size these models are to each other. Is Fable roughly equivalent in size to Astra? Opus roughly equivalent in size to Sol? Is Neon more similar to Fable or Opus?

    • elxr 3 hours ago

      Carbon/Argon/Hydrogen - Just skip. Simple.

      Neon - Probably decent based on twitter impressions and benchmarks.

      For anthropic, just listen to anybody that heavily codes agentically, it should be clear which model is the best for most typical tasks (Opus).

      OpenAI though, it's definitely less clear right now. I haven't been enjoying either 6.2 astra or 6.2 sol, and they're definitely occasionally bone-headed in a way I haven't seen before on a frontier model this late into the game. If the name is the biggest of your concerns, you're not doing enough work.

    • floatingtorch 4 hours ago

      I wish they would all just use huge, big, small or something like a clear ordering. Similar reasoning levels which are very intuitive.

      • MrDunham 2 hours ago

        Agreed... though at some point I wonder if they paint themself into a corner with that...

        "Even Huger", "Most Hugest", and "Most-Most-Hugest-v2_Final_Final" feels like a real possibility.

        Fable is a good example. "It's Mythos (real big)... but Mythos-light, so we'll call it Fable". Which is smaller than an Opus.

    • rtkwe 5 hours ago

      Yeah I don't even try, whatever Copilot sends it to seems to work well enough for me at the moment.

    • robofanatic 5 hours ago

      reminded me of Android dessert version names Cupcake, Donut, Gingerbread, Ice cream sandwich….

      • SoMomentary 4 hours ago

        IIRC those at least did us the courtesy of proceeding alphabetically.

    • B-Con 5 hours ago

      I'm terrible with names, so I'm in the same boat. There's a lot of marketing going on.

      But the naming isn't completely random, here's what I use to make the ordered list in my head:

      Claude is named after creative works, and uses bigger works for more powerful models. In order of strength:

      * Haiku - short poem (we all know this one)

      * Sonnet - "medium" length poem (many of us know this one)

      * Opus - It's the singular form of "opera", so it's like a stand-alone segment of a larger body of work. (Nobody knew this one, but not hard to remember, and know you know that "opera" is actually plural! The description sounds bigger than a sonnet, but it's not a strict ordering.)

      * Fable - a story (this one doesn't seem obviously bigger than an opus, but I guess an opus is one part and a fable is a front-to-end story, it works)

      * Mythos - another type of story, which works and it's basically Fable but Anthropic's internal-only version

      Gemini, as of version 4, started doing something with atoms:

      * Hydrogen - smallest periodic element, the most common gas in outer space

      * Neon - a bigger element, less common in outer space

      * Argon - yet bigger element, yet less common in outer space (as best I can tell, I am not an astronomer)

      * Carbon - this breaks that pattern, its a smaller element than the first three, and not a gas... but critical for life on this planet? And it sounds like a longer version of Argon. So I can remember it, but I'm not sure what the intended pattern was.

      The model ordering is not strictly upgrades, but AFAIK they are intended to be steps up in terms of (claimed) complexity and reasoning power. But since it's an evolving field, the specifics of how cost, speed, and accuracy combine can make it hard to know how to order a given model for your use case. (i.e. cheap & fast with a harness that makes more calls, or expensive and slow but better at one-shot).

      Deepseek's naming is more organized and literal, but you still have to remember the prefixes for V = version (plain) and R = Reasoning, and that they have different specialized models that aren't intended to be compared to each other.

      • adrianmonk 4 hours ago

        I'm not sure if it's intentional, but the Google AI code names are all elements used in producing light.

        * Neon and Argon: noble gases commonly used in gas-discharge lamps[1].

        * Carbon: used in arc lamps[2] (the first practical form of artificial electric light).

        * Hydrogen: the sun (abundant light, available to everyone).

        The symbolism of light would make sense for AI. Shedding light on a subject, enlightenment, a light bulb moment, etc.

        If this is right, then maybe we'll see krypton, xenon, mercury, sodium, or tungsten next.

        ---

        [1] https://en.wikipedia.org/wiki/Gas-discharge_lamp

        [2] https://en.wikipedia.org/wiki/Arc_lamp

      • surround 5 hours ago

        Basically every lab has a model that doesn't fit their naming scheme 100%.

        OpenAI's naming scheme is in Latin:

        * Luna - Moon (smallest)

        * Terra - Earth (medium)

        * Sol - Sun (large)

        * Astra - Star (could be larger than the sun I guess)

        Meta's Muse:

        * Ember - (smallest)

        * Spark - (medium)

        * Wonder - (large)

        A spark is the beginning of an idea whereas wonder is associated deep thought. But an ember and a spark are both small pieces of fire, so the ordering is a little muddy.

        Google's naming scheme has the least clear pattern of any lab.

        • mmoskal 4 hours ago

          Astra is actually stars plural. One is "astrum" apparently.

        • bee_rider 4 hours ago

          I wonder if it was supposed to grow in fire size. Ember and spark fit the pattern and there’s plenty of room to grow. But eventually they’d start reaching sizes related to catastrophes.

          • Aeolun 4 hours ago

            Muse Wildfire, Muse Conflagration

        • tintor 5 hours ago

          Yes, it is Astra because 'GPT 6 Black Hole' didn't get approved by marketing department.

          • zzzoom 4 hours ago

            And people don't know Sagittarius A*

          • proton_9 4 hours ago

            haha! this made me chuckle, perhaps they release black hole when they can claim to have achieved singularity

      • throwaway473825 5 hours ago

        Speaking of Opus, many have likely heard of the expression "magnum opus" (Latin for great work). That's probably easier to remember than the singular form of opera.

      • fimi 5 hours ago

        At first I thought that has lexicographic order, like Argon is small, Carbon is big, and something between.. but then Hydrogen, and Neon, whatever

      • LoganDark 5 hours ago

        Mythos and Fable are identical now except for "vetted organizations". I'm not even sure why they're still offered separately in the CVP.

    • crazygringo 5 hours ago

      Seriously. It's annoying enough to keep track of the names of which models mean what just within Claude Code. But they don't change that often, so I can deal.

      But trying to keep track of the model names in all 3 companies? Utter madness.

      We need something better. All I really care about most of the time is whether I'm using the cheapo, average, or expensive model. I don't really care about which actual model is occupying which actual spot at which time with which cutesy name. I just know I have a budget, and so I allocate the three levels according to the type of work I'm doing.

    • asaddhamani an hour ago

      We had names like nano, mini, flash, flash-lite, pro, etc. These were easy to understand. I have no idea what’s going on with this helium hydrogen carbon germanium nonsense. Do they want me to take out a copy of the periodic table to understand how the models compare? Is that even how it works? Flash lite, flash, pro wasn’t broken. Why did they “fix” it?

    • nl 3 hours ago

      I think vendors assume their names schemes are obvious.

      Anthropic: "We'll name after the size of literature style! An opus is obviously bigger than a sonnet!"

      Much of the world: "What is this word Opus?"

      OpenAI: "We'll use the Latin names for astronomical bodies. Everyone knows that terra = earth and luna = the moon and that the earth is bigger than the moon."

      Much of the world: "Does terra mean it is terrified? Oh well at least it is better than Gpt 4o vs Gpt o3"

      Google: "We'll name it Bard! Oh, that was bad. Lets name it Gemini because that sounds cool. Oh we need different models, ok we'll have Pro for the big expensive on, and Flash for the fast one. But now we want a fast large one and also we have mid sized models. Ok, lets name them after elements, but not ones that anyone knows anything about so now we can confuse them as to if Neon or Carbon or Argon is better. But Gemini Hydrogen is the cheap fast one"

      Everyone: "Ok, Gemini Hydrogen is cheap. But this is too confusing, I'll just use Claude".

    • datadrivenangel 4 hours ago
    • willis936 7 hours ago

      Maybe it's the same marketing direction that landed starbucks' drink size names.

      • virtue3 7 hours ago

        Venti is just Italian for 20 for 20oz

        Yeah grande(big?) and tall are whatever to your point

      • colejohnson66 7 hours ago

        At least Starbucks doesn't change their drink size names every few months.

    • com2kid 3 hours ago

      If you ask Gemini what it thinks about Gemini model names even it'll agree it is a bad product branding strategy.

      So, there is that...

      Luna, Tera, Sol, that makes sense. Size. Yay. Big thing better.

    • Art9681 2 hours ago

      The naming scheme got you to take precious time and talk about it. All marketing is good marketing.

      I still agree with you though.

    • eminence32 7 hours ago

      It doesn't help that the blog post opens with a long description of all the benchmarks that the model excels on, but very little info about how I should pick what model to use.

      I do note that gemini.google.com has replaced Argon with Neon, so I guess maybe that's a clue?

      My personal approach is to not really care at all. These specific models aren't going to be here in 6 months, so learning the intricates of these specific models has no benefit for me. I'm just going to mostly rely on whatever one is the default

    • ionwake 5 hours ago

      Just use animals guys. Predator prey for thinking, nimble more reactionary etc.

      Expert models: Lion, Shark, Whale

      Flash models: Swift, Falcon, Robin

      Though I still think it shoudl jsut be : Expert v13 , Fast v3

      • nimbleal an hour ago

        From a user’s point of view I think it really comes down to the old “fast cheap powerful — you can only pick two”. I’d love

        Fast and cheap Fast and powerful Cheap and powerful Kinda all three but not really

        Not very catchy names though

    • quixoticelixer- 6 hours ago

      It's possible they are just making it confusing as possible so everyone switches to the orchestrator but I bet they are just as confused as the rest of us.

    • frogperson an hour ago

      i really cant believe there isn't a curated type of "app store" for popular models, built for people who don't live and breath this stuff. Why cant i type in a use case and get a model recommendation? Why is there no up/down voting for models? surely someone could slop that together with a catchy name.

    • coolestguy 2 hours ago

      It's completely ridiculous when not even tech people can keep up - how can a normal person?? Just name them, small/medium/large then a revision number

    • rao-v 5 hours ago

      Atleast there is some hope we'll get Gemini 5.3 Arsenic one day

      • tintor 5 hours ago

        Gemini Lead, Gemini Uranium, Gemini Plutonium

    • pizzathyme 7 hours ago

      I thought the higher periodic table numbers meant a more powerful capable model. But not even that appears to be the case

    • j9feng 5 hours ago

      agreed! Apparently their AI is not intelligent enough to pick proper name schemes for itself.

    • paulddraper 6 hours ago

      What do Corolla, Camry, Crown mean?

      While cutesy, most current models are quasi-semantic progressions.

      Claude (length, or cultural durability): Haiku, Sonnet, Opus, Fable

      GPT (altitude): Terra, Luna, Sol, Astra

      ---

      Gemini unfortunately has no semantic key Hydrogen, Carbon, Argon (superseded by Carbon), Neon.

      • motionlessveloc 6 hours ago

        Corolla = Corona = Latin for "Crown"

        Camry = Kammuri = Japanese for "Crown"

        Crown = English for "Crown"

        • redwall_hp 4 hours ago

          There was also a Corona until 2001, and briefly a Tiara in the 60s.

        • paulddraper 4 hours ago

          That makes sense why Toyota replaced “Avalon” with “Crown”.

      • stillpointlab 4 hours ago

        I thought about this. Does that mean renting a model is like leasing a car?

        Anthropic is the new Mercedes? OpenAI BMW?

      • j0s4 3 hours ago

        You have the GPT ordering wrong though, as Terra is generally more capable than Luna

      • happyopossum 5 hours ago

        > What do Corolla, Camry, Crown mean?

        This wouldn't be a conversation if the big 3 didn't change their model names multiple times per year. Or conversely - do you think people would be fine with Toyota renaming the Camry every 6-9 months?

    • closetheloopdev 7 hours ago

      Maybe the trick to working with these RaNd names is for us to tell Claude Fable/Opus or GPT Astra/Sol: "go find the Gemini model that is as smart as you and discuss XYZ with it, and then tell me what you guys agree on."

      I feel like Google purposely names their models not for humans but for AI agents. People keep saying: "make your products ready for agents," so maybe that's what Google is doing, which is making Gemini 4 models ready for other agents to use.

    • oofbey 7 hours ago

      I'm ready for them to just issue UUIDs for each release and use those as the marketing names. At least it's honest.

    • qwerpy 7 hours ago

      Someone go and vibe code a levels.fyi type thing but for LLMs instead of SDE compensation.

      • fritzo 7 hours ago

        Wait, where the LLMs write how it feels to work inside different companies?

    • ijidak 2 hours ago

      Well Anthropic and OpenAI are actually doing a great job with their naming.

      We intuitively understand that Luna is smaller than Terra, and Terra is smaller than Sol, and Sol is smaller than Astra.

      Likewise, we intuitively understand that a haiku is shorter than a sonnet, and a sonnet is shorter than an opus, and an opus is shorter than a fable. Fable is probably the weakest of their names. Maybe they should have gone with Epic.

      It's Google that isn't doing it right.

      Even the educated have no intuition regarding carbon, argon, and neon.

      Hydrogen being the smallest makes sense.

      But what property is the point of comparison for carbon argon and neon?

      Valence shell? Atomic mass?

      It provides no self-evident point of comparison. It's absolutely atrocious.

    • jstummbillig 7 hours ago

      That's hilariously banal for a top comment on a model release. Maybe it speaks to the banality of model releases.

      • hadlock 7 hours ago

        We've moved to <family> <variant> <number> e.g. deepseek flash 4.1. I can keep up with that, and if/when Flash 4.2 comes out, great. I can grok that. Google's naming convention just shows how out of touch management is with the market.

        Android's silly food naming convention works when you do one release every ~18 months and you are the market leader. It makes zero sense when you're releasing 8 models a year

      • djmips 5 hours ago

        Yeah they could just use AI to translate the model names and tiers to whatever floats their boat.

    • dfdeath 6 hours ago

      Clicked this expecting the top comment to be a gripe about naming conventions, was not disappointed. But the trope is getting a little old at this point. It might be time to consider another irrelevant talking point you can beat to death.

    • catigula 7 hours ago

      The OpenAI and Claude naming schemes are fairly easy to discern with no prior knowledge of the models.

      • Fabricio20 7 hours ago

        I dont agree that Anthropics names are easy for an outsider, i'd have no clue what a Fable is compared to Haiku, to me both are just "story" names. OpenAI's new naming scheme is celestial bodies and i'd assume it's just size, bigger celestial body bigger performance, except i dont know where "Astra" fits since Sol > Terra > Luna at least make sense. Now gemini... I was just thinking it today that "Carbon" and "Hydrogen" with hydrogen being the lightest element? Is that what they are going for? I wouldn't know off the top of my head where Neon and Argon are placed though.

        Edit: Just as I post this someone else comments google's doesnt follow elemental sizes at all so my theory is wrong too!

        • WaltPurvis 6 hours ago

          I'm obviously thick, but this is the first time it's dawned on me that the Claude names were types of written work. Now I'm wondering why Opus isn't the top model; isn't an opus a more impressive (or at least larger) piece of work than a fable?

          • kubik369 6 hours ago

            I don't have a cohesive explanation for each one, but in my mind a fable is basically a legend (or in other words, a myth/mythos) and a legend is the final form of a story. Even though it has been half forgotten and changed, it still lives on, affecting people across times greater than a human's short life.

            • lucumo 4 hours ago

              A fable is just a story with antropomorphized animals and a moral. They can get very short.

              Here's a translation of the entire Fox and Grapes fable:

              > Driven by hunger, a fox tried to reach some grapes hanging high on the vine but was unable to, although he leaped with all his strength. As he went away, the fox remarked "Oh, you aren't even ripe yet! I don't need any sour grapes." People who speak disparagingly of things that they cannot attain would do well to apply this story to themselves.

          • atombender 6 hours ago

            An opus is just a musical composition, sometimes also used about written work, but it doesn't really imply a great work. Maybe you're thinking of "magnum opus"?

        • hrncn 7 hours ago

          The point is that Haiku was Haiku a year ago, and the number keeps increasing, while googles gemnini started with sane names and numbers as 1,2,3.x pro/flash/lite, and is now suddenly continuing with 4 hydrogen/argon/neon/carbon something.

      • stillpointlab 7 hours ago

        For me, it is familiarity that does the work. I'm familiar with Sonnet, Opus and Fable and what they mean to me. The same is true for Sol and Astra bu I'm not sure about Terra or Luna.

        So I'm just hoping for consistency over some period of time to allow me to get familiar.

      • bogdan 7 hours ago

        No way Claude is obvious without prior knowledge

        • hannasanarion 6 hours ago

          Prior knowledge of common poetic forms taught in middle school?

          The point is that there is prior knowledge that completely explains the entire system. A 3 line poem < a 14 line poem < a book-sized work < a multi-book epic, boom done, easy. Consistent and friendly using words that everyone should know.

          With google we have element 1 < element 18 < element 10 < element 6. There is no consistent schema where their namesakes map to their relative sizes, price, and capability.

          • koe123 5 hours ago

            While I agree that Anthropics naming makes more sense, I also doubt most people know the periodic table that well

          • happyopossum 5 hours ago

            > A 3 line poem < a 14 line poem < a book-sized work < a multi-book epic

            So wait, which is smarter - Fable or Opus? It's clearly not obvious to most people or this debate wouldn't be happening.

            As the old saying goes - if you're explaining, you're losing.

          • DylanBohlender 6 hours ago

            Valence electron counts don't even work either. The noble gases are in the middle! It's insane!

          • staticman2 6 hours ago

            You are now claiming a Fable is a poetic form?

      • jabron 7 hours ago

        They still require prior knowledge of Latin words of celestial bodies or different words for literary works, same as Google's requires knowledge of the periodic table.

        • nicoburns 7 hours ago

          > Google's requires knowledge of the periodic table

          The problem with Google's is that knowledge of the periodic table doesn't seem to help.

        • nater5000 7 hours ago

          Damn, if only the people concerned with what the naming of AI models mean had access to some magical technology they could use to ask what the names of those models mean.

        • yunwal 7 hours ago

          What would knowing the periodic table tell you about google's models? They are not in order of atomic units! Hydrogen (1) < Neon (10) < Carbon (6)

          • mmahemoff 7 hours ago

            They could have used a generative “*gen” theme with some of the most fundamental elements: Hydrogen, Nitrogen, Oxygen.

          • hannasanarion 6 hours ago

            Don't forget last month's Argon (18) which is better than Hydrogen (1) but worse than Neon (10).

          • dezgeg 7 hours ago

            ASCIIbetical order, obviously!

      • yard2010 37 minutes ago

        Right it's so easy - luna is the best, and davinci-002 is the top. Everything else is in between

      • jbjbjbjb 7 hours ago

        The number of replies arguing that the OpenAI and Claude naming is confusing doesn’t bode well for human intelligence

      • carterpaul 7 hours ago

        I mean, they're OK, still a bit cryptic. Astra in particular is a bit ambiguous - they're all named after celestial bodies then the largest is just named after space in general.

        Would love "small", "medium", "large", "extra-large", etc...

        • Nesco 7 hours ago

          mistral does this

        • ascorbic 7 hours ago

          Astra means "stars"

          • alanfranz 7 hours ago

            The Sun is a star itself. More stars are better? What about white dwarfs?

            Systema would have worked. And maybe galaxias after that.

            • spolitry 6 hours ago

              Astra is basically all of systema and galaxias

        • _delirium 7 hours ago

          Google used to do that with a version number and descriptor. 2.5 Pro, 3.1 Flash-Lite, etc. Gave you both model generation + idea of relative size/quantization within generation. I guess they decided that scheme wasn't flashy enough?

        • Brendinooo 5 hours ago

          Having named a range of sizes more than few times by now...S/M/L/XL always starts out great and then turns into a MESS of XS, XXXL, Medium-Large, forever and ever.

          • lucumo 4 hours ago

            Yeah, or you get an escalatory scale. Similar to how LF/MF/HF (low/medium/high frequency) got "very HF", "ultra HF", "super HF", "extremely HF", "tremendously HF".

  • bisonbear 8 hours ago

    Anthropic competing with Google instead of OpenAI at the frontier is unexpected, I have been doubtful of Google DeepMind's status as a frontier lab in the past few months, but damn they've been cooking recently

    • Tadpole9181 7 hours ago

      It's important to note that DeepMind and the other two labs aren't playing the same game.

      The latter need to raise funds and IPO, but the former is an arm of one of the most powerful companies to ever exist with some of the largest data warehouses around and custom hardware. And they have an endless pile of internal code to maintain already with hundreds of millions of end users already bought-in.

      The incentives being different leads to different focuses and product lifecycles.

  • jaykru 8 hours ago

    Carbon 4 has been incredible for my use, tuning a local inference engine based on tinygrad running on Tenstorrent hardware. I had a barely functional prototype I built with Claude that petered out at an unusable level of performance. Astra and Sol didn't make much progress when I pointed them at it. Carbon 4 more than 10x'd the prefill and doubled the decode performance on Qwen 3.8 27B without breaking quality or the generality of the stack. Very nice to see competition at the frontier again.

    • henrymerrilees 3 hours ago

      This sounds like a blast!

    • blueblisters 2 hours ago

      A plausible explanation here is that Ant and OAI are deliberately sandbagging on ML / GPU optimization tasks to preserve their lead while Google doesn’t give a fuck.

      Argon is incredible at AI research and its responses are grokkable

      • jaykru an hour ago

        It's also really good at my fuzzier math/research questions :shrug: more insightful than fable imo. The nerfing is real (indeed Ant at least admit to it) but I don't really feel like it's a useful fact to have in mind when evaluating the products I pay for.

    • atomashevic 8 hours ago

      Just curious, what TT hardware do you have? How many tps did you achieve with Qwen 3.8 27B?

      • jaykru 7 hours ago

        I'm running 2x p150a's right now. I have them hooked up to a mini-ITX PC with a single PCIe slot via a quad x1 switch (figuring out power to wire up the other two cards I own). My best numbers so far are around 2500t/s prefill about 52t/s decode with DFlash2. All of this is vibecoded besides the original design sketch which was maybe a 3 line description of what tweaks to make to tinygrad and which parts of the existing TT stack to use (almost nothing).

  • irthomasthomas 6 hours ago

    Looks like it gets to within about 1% of Fable 5.2's performance for about 40% the cost on vals.ai https://www.vals.ai/benchmarks/vals_index

  • numpad0 8 hours ago

    Hydrogen < Neon < Carbon? But where's Argon? Why those elements?

      01 02 ... 13 14 15 16 17 18 
      H                        He
      Li Be ...  B  C  N  O  F Ne
      Na Mg ... Al Si  P  S Cl Ar
    
    ???
    • wernam 8 hours ago

      Maybe Knight moves on the periodic table :)

    • seizethecheese 7 hours ago

      I got tripped up assuming bigger was a larger element. Seems like it's vibe, like Carbon feels bigger than Neon.

    • ajs1998 8 hours ago

      There is no logic. Someone made a list of cool element names and when they run out they'll switch to some other scheme that communicates nothing

      • thehappypm 5 hours ago

        It’s alphabetical chronologically

        • deaux 3 hours ago

          The fact that there's a dozen comments on this post alone talking about periodic tables, element sizes and other stuff and these are the top comments and you and me are seemingly the only ones noticing this pretty obvious pattern..

          It's not like this has no precedence. Google themselves are the best known example in software with Android codenames. Starting with [A]rgon was a clear signal in and of itself. But above all, just basic common sense strongly goes against Google uses some very limiting, niche pattern rather than just alphabetic.

          I'd chalk it up to eternal September but I'm not convinced most of these people are new. A particular brand of neurodiversity, I guess.

  • chriddyp 6 hours ago

    Just ran my data analytics benchmark [1] on the model: Excellent results for this model like the rest in the Gemini 4 series (all models score an A).

    Costs spread between Hydrogen ($6) < Carbon ($11) < Argon ($13) < Neon ($17) for this 43 question exam.

    At a quick glance, the tone is good as well. But we do a fair bit of prompting in our harness to guide the model, so I can't say how the default model behaves.

    [1] https://plotly.com/blog/gemini-4-neon-plotly-data-analytics-...

  • xnx 8 hours ago

    > Neon writes clearly, with well-structured responses and a natural, direct tone

    This is an underrated quality of Gemini models.

    • Tadpole9181 7 hours ago

      Does anybody have a corpus of examples I could look at? I've wondered if it's worth setting up Gemini as a human-facing documentation author with other models still being the underlying workhorse.

  • brap 8 hours ago

    Google is on a roll, what is this the 4th release this month?

    • nateb2022 8 hours ago

      RSI

      • brap 8 hours ago

        Sure looks like it

        • catigula 8 hours ago

          It looks like a distilled model.

      • inknight 8 hours ago

        Please have them recursively improve their model names

        • mpvjs 8 hours ago

          - please improve the model names

          ChatGPT thinking: the user is not happy with my previous solution, propose a cooler scheme like toy story character names...

    • tanx16 8 hours ago

      They're releasing... periodically

  • jrflo 8 hours ago

    Between Astra and Fable on the Pareto frontier:

    https://artificialanalysis.ai/#intelligence-efficiency-tabs

    • jeffbee 8 hours ago

      That looks to me like OpenAI equals or beats everything Google offers at half the price.

      • jrflo 8 hours ago

        You're right, it's 1-2 points shy of the frontier

        • jeffbee 3 hours ago

          I'm surprised they can't afford to just slide the prices over to the left and eat the losses.

  • Johnny_Bonk 3 hours ago

    If theres one thing id like to see from ai development, some third party verified certifications as to how much usage each plan gives you… i can barely try it on antigravity pro let alone know what the other plans will give and its always just a multiplier and some quantizing going on in the background etc

  • s20n 3 hours ago

    They also seem to have removed Argon from the gemini app and replaced it with Neon. I guess Argon was just too inefficient even for google that they had to pull the plug on it so fast.

    • Andrex 2 hours ago

      I became accustomed to my Argon queries eating 14-28% of my hourly capacity each prompt.

  • xd1936 8 hours ago

    I have the "Google AI Pro 5TB" plan for Students. I see "4 Hydrogen" and "4 Neon" in the model picker on gemini.google.com. When I look at one.google.com's plan details, I see "Your Plan: Google AI Pro (5 TB)". When I click "View Benefits" on that page, one option in the list says:

    Google Antigravity

    Entry rate limits to advanced models, including Gemini 4 Argon and Opus 5.5, in Google Antigravity, our agentic development platform

    Yet, here I am, running the `agy` cli, and the model picker says "Unavailable: Gemini 4 Carbon/Argon/Hydrogen/Neon is currently available to paid Pro and Ultra subscribers". It has been this way since the release of Carbon. I guess it's "3.8 Flash" or nothing for me.

    • mathis 6 hours ago

      Hydrogen, Argon and Neon are included in Pro. Carbon is not. At least for me. You might want to update your cli.

      • pzo 6 hours ago

        Not true. Hydrogen and Neon are available only in web chat. 3.8 Flash is the only one available in antigravity for such promo subscriptions.

        • rawgreaze 4 hours ago

          I had to grab my laptop to say, you are actually incorrect. I have been using Argon for the last week or two via Antigravity on the Google Pro $20/mo plan, discounted to $5/mo for 3 months, just fine. And as of today yes, I see Neon and Hydrogen there. The only one I can't see/use is Carbon. https://i.imgur.com/2c9EnbV.png. This isn't just a visual thing, I have actually used every model there. They all work.

  • aizk 8 hours ago

    Google is back in action, but the real question is why is OpenAI falling behind, seemingly.

    • kccqzy 3 hours ago

      They paused on new model development because their old models were routinely escaping their terribly weak sandbox to wreck havoc in the real world during evals.

  • artdigital 8 hours ago

    Okay so it seems like the ranking here is Hydrogen -> Neon -> Carbon even though that is not clear at all from this release. At least Googles model picker in the Gemini app now lists those 3 above each other in this order, and keeps Carbon for “Code & complex work”

    So Neon is the new default, and Carbon the slow deep thinking “when you need it”? A bit like how Opus 5.5 was better than Fable 5.1 in most things, but Fable still had the edge over deeper things (then 5.2 came along)

    Hydrogen has already became my default for 80% of small tasks. If Neon is as fast but smarter that would honestly be very cool

    • vtail 8 hours ago

      Do you know what happened to Argon? Was it a flop?

      • shric 8 hours ago

        For some reason nobody reacted to it

      • artdigital 7 hours ago

        It was good, but Carbon quickly superseded it

      • lukax 8 hours ago

        Carbon is Argon with extended thinking

    • gradus_ad 8 hours ago

      To be fair prior to Gemini 4, the app had 3.1 as the "Complex work" model while 3.8 Flash, far superior in every way to 3.1, was listed as the "Fast" model.

      So I'd take their hierarchy with a grain of Sodium.

  • horseloverthin 8 hours ago

    I think that its likely the reason for a) the strange naming scheme b) the promptness between releases is that google has outpaced internal expectations for model development ( probably due to RSI ) and now can now utilize its vast compute pool much more effectively for autoresearch and exploratory self-improvement tasks. I wonder if this suggests a steeper trajectory for them than OAI/Anthropic in general.

  • stbenjam 6 hours ago

    Probably the best bowls of ramen I've seen from a Gemini model

    https://bitbin.de/ramen-bench/#google%2Fgemini-4-neon%2Fxhig...

  • mertarican 7 hours ago

    Well, I use my product as a benchmark for this new models and I have to say that Argon is still the most efficient model by far for my use case. Carbon, Hydrogen, and now Neon all use 3 to 4 times higher tokens compared to Argon for creating demos. Neon was the first Google model that passed 8192 MAX OUTPUT TOKEN limit, which is disappointing a little bit. But lot's of experience I got from these models, for example bigger models like Argon spend significantly less output token so even though unit cost is higher both total cost and the output time can be lower than the lower models.

  • maplet 8 hours ago

    > Independent evaluations rank Neon as one of the top models and on benchmarks like the Vals Index, it ranks #2 (70.9%), and reaches #1 on both the Legal Research Bench and Tax Agent Bench.

    On Artificial Analysis Intelligence, it scores comparably to GPT-6.2 Sol but is significantly more expensive.

  • bluegatty 8 hours ago

    There are too many names and versions to keep up with.

    They all need to pick a dam naming scheme and stay with it.

  • raincole 8 hours ago

    It looks like argon is deprecated.

    • doodlesdev 8 hours ago

      Already removed from the Gemini app. Replaced by Neon.

    • vintermann 9 minutes ago

      That was quick. Really hope this new one hasn't regressed on historical HTR.

  • XCSme 6 hours ago

    In my tests it's a lot more expensive than 6.2 Sol (if used at max ressoning level, but medium results in similar score).

    Gemini 4 Argin vs 6.2 Sol vs 6.2 Astra vs Fable 5.2: https://aibenchy.com/compare/google-gemini-4-neon-xhigh/open...

  • epsteingpt 4 hours ago

    It's a very good model. If they get their UX right, it has capabilities that Opus 5.5 and Fable could never have (e.g. watching YouTube videos, generating images).

    Race is getting very interesting again.

  • mvid 5 hours ago

    Google models may as well not exist, since they can't be used outside of the official harnesses

  • vektormemory 5 hours ago

    Seems like they are finally on a roll.

    Eventually all these models will have to merge into a Omni model with image/video/tts voice/text all in one.

    You try keeping 600+ models in our tables updated from all the model providers; it's a part-time job every day.

  • lanthissa 8 hours ago

    looks good gemini team fixed the massive spread they had vs openai and anthropic in terminal bench - swe bench.

    That spread is my personal bs detector for models over promising.

    • andriy_koval 7 hours ago

      I suspect all labs just benchmaxed those benches by now.

  • woggy 7 hours ago

    I don't get it, why is Neon better than Carbon in the benchmarks but then in the end they say Carbon is better?

    • chicagobuss 5 hours ago

      it's like how sonnet was trading blows with opus and opus 5.5 was ahead of fable 5 for a while

  • agrippanux 8 hours ago

    All I really want now from any provider is Opus 5.5-level at blistering speed and a fraction of the cost. How close does this come?

  • dennysora-main 8 hours ago

    This is good, it makes me have to understand the full names of every element in the periodic table.\ (PS. Taiwanese)

    But I still can't understand the differences between these models.

    Recently I watched Veritasium talking about helium, and I was thinking, will the effort setting of these models eventually have something like "Temperature K" to configure?

  • darkotic 3 hours ago

    Possibly Gas/Flash models; Hydrogen, Neon, Argon Solid/Premier models; Carbon

  • abraxas 8 hours ago

    Is it true that this one has a 1MM token OUTPUT window unlike the other competing models that are capped at 128K?

    • Computer0 8 hours ago

      With agents often taking many turns before responding to you, what use case does this change for you?

      • abraxas 8 hours ago

        I think for large refactoring jobs this will open up a new category of work that an agent will be able to do. Also reasoning steps count towards output tokens, no?

        • Computer0 7 hours ago

          That is a good point, I hear of sonnet frequently hitting a limit in its thinking.

  • tech-historian 7 hours ago

    Andon Labs says Gemini 4 Neon is #1 on Vending-Bench 2.

    https://x.com/andonlabs/status/2123174738789011875

  • fooblaster 2 hours ago

    Has anyone tried this?

  • bpiroman 5 hours ago

    New AI models are kind of like new iPhone releases. Looks cool but not much different to the previous model.

  • low_tech_punk 8 hours ago

    I'm so confused now. Carbon and Argon already have overlaps, and they are adding Neon to the mix, with almost the same price and marginal improvement.

  • gizmodo59 7 hours ago

    I did not know there is a bench called reinmann bench. So does this mean they are closer to proving/disproving that conjecture?

    • Destructotor 5 hours ago

      Looks to me like just another math benchmark similar to FrontierMath.

      > We introduce Riemann-Bench, a private benchmark of expert-curated problems designed to evaluate AI systems on research-level mathematics that goes far beyond the olympiad frontier. Problems are authored by Ivy League mathematics professors, graduate students, and PhD-holding IMO medalists, and routinely took their authors weeks to solve independently. Each problem undergoes double-blind verification by two independent domain experts who must solve the problem from scratch, and yields a unique, closed-form solution assessed by programmatic verifiers.

      https://arxiv.org/abs/2604.06802

  • dprkh 8 hours ago

    Has anyone else been having an issue where the models would answer its own question? I have been seeing this quite a lot with Gemini 4 models.

    • smallerize 8 hours ago

      The "long-horizon" ones are optimized for this. They won't get a high score on the benchmarks if they stop to ask questions.

      • vintermann 8 minutes ago

        And I hate it.

      • dprkh 7 hours ago

        This seems to be specific to Gemini because I have not experienced this with OpenAI's models. In fact, Codex now has a feature such that the model can ask a question and continue making progress on work that is not blocked while waiting on your answer.

        • smallerize 4 hours ago

          I wasn't benchmarking or anything rigorous, but I noticed it with GLM 5.3 Flash. It was downloading and installing stuff to my computer to try to work around an assumption in the prompt that it knew was wrong. I also noticed it trying to work around websites' bot prevention measures. I stopped it often.

    • jeffbee 8 hours ago

      Oh yeah, and you're paying for it to say "wait. what about [[thing]]?" then answer itself.

  • mrbungie 8 hours ago

    Is this another release that I can't use with agy? (I totally refuse to pay 100 USD just to try it in that context)

    Edit: I stand corrected, it can be used in agy.

    Edit2: I have an student account and students accounts can't use these models apparently, https://discuss.ai.google.dev/t/bug-entitlement-sync-paid-go...

    • anvuong 8 hours ago

      It's there in my $20 sub. Only Carbon needs $100+

    • lanthissa 8 hours ago

      you can use this with agy, im on pro and just tried it it has its own orchestrator too for delegating to hydrogen.

      • mrbungie 8 hours ago

        Oh cool, I'll check it then. Last time I checked I only had 3.8 flash and 3.1 pro.

  • BoldColdHold 8 hours ago

    Ok this is a good model by the looks of it. Mid size model that only loses to fable.

  • jonathanstrange 8 hours ago

    They have just removed 3.1 Pro and Argon (and Flash) from my Pro subscription without any warning. I was using 3.1 Pro for some important workflows. It was actually better than the newer models for some tasks. Now there is suddenly only "Auto", "Hydrogen", and "Neon"?

    Maybe the people at Google don't realize that but some people need constancy. If a model was good enough for some work 6 months ago, it's not suddenly becoming not good enough for the same task.

  • sergiotapia 2 hours ago

    I have never used any google models before yet I hear this Neon is tremendously powerful. What would be the best way to use Neon? How do you use Neon (or other google models) - do you need to use antigravity?

    https://x.com/LH/status/2123225395298537974

  • ariwilson 8 hours ago

    so which model am I supposed to use for what? someone help me

    and why can't I use any of these with agy and a $20 plan? unlike openai and anthropic

  • holistio 7 hours ago

    I know it's now trendy to crap on Gemini, but Hydrogen was a very pleasant surprise after having used it for various personal tasks.

    I'm curious about Neon.

  • pilooch 8 hours ago

    So this is above carbon ?

  • tonyhart7 2 hours ago

    they release a lot of model so fast, do they really achieve RSI ????

  • 3wolf 8 hours ago

    Credit to Google for outdoing Open AI and Anthropic in terms of choosing a confusing naming convention.

    • xattt 8 hours ago

      Is it not going by atomic mass?

  • XCSme 8 hours ago

    So it's better and cheaper than the recently released Carbon?

    So Carbon is already deprecated, or?

    EDIT: They do explain where it sits in the line-up here:

    https://blog.google/innovation-and-ai/models-and-research/ge...

  • psidebot an hour ago

    For a significant product announcement, couldn't they have at least gone for a real neon sign? I mean AI imagery has come a long way in a lot of areas, but their announcement image is seriously uncanny valley lame.

  • WarmWash 8 hours ago

    Neon is the long awaited successor to 3.1 pro. Took a while to get there

    So now we have

    (XL): Carbon/Argon Astra Fable

    (L): Neon Sol Opus

    (M): Hydrogen Terra Sonnet

    (S): Helix(rumored) Luna Haiku

    • agentdev001 3 hours ago

      Yea- if we're just talking size these would be the tiers, but if you're measuring by API spend then things certainly get messy.

    • 5555watch 8 hours ago

      I don't believe Argon is XL level??

      Also, very curious whether they will ever release the rumored Krypton model, or will it be hidden ala Mythos

      • WarmWash 8 hours ago

        Argon is a big model, but it was completed sometime in mid-august, making it kind of old in LLM land.

    • Melatonic 8 hours ago

      Other people are saying that Neon replaced Argon?

      • WarmWash 8 hours ago

        It very well might be better. Nowadays being newer tends to matter more than being bigger.

    • oblio 2 hours ago

      > Neon is the long awaited successor to 3.1 pro. Took a while to get there.

      The real question is: will it be available for Google Workspace? Not everyone wants to pay for a dedicated AI license :-(

  • nandanadileep29 7 minutes ago

    better than opus 5.5? hmm

  • minimumvector 3 hours ago

    On AA's Intelligence Index (https://artificialanalysis.ai/), the ranking is Fable > Astra > Neon > Opus > Sol.

    On Vals's Index (https://www.vals.ai/benchmarks/vals_index), it's Fable > Neon > Sonnet > Opus > ... > Astra.

    On Surge's Tuesday@Work Index (https://surgehq.ai/benchmarks), it's Neon > Fable > Astra > Opus > Sol.

    Overall, it seems like Gemini is back at the frontier with Anthropic and OpenAI, and well ahead of Meta and Grok. Despite what folks were claiming a couple months ago - when after Jeff Dean and Noam Shazeer left, they were saying Gemini would never be SOTA again.

    Once Anthropic and OpenAI are public, I wonder how new releases like this will affect their stock prices?

  • gradus_ad 8 hours ago

    So Anthropic is being valued at around $1.2T and their only revenue stream is a collection of frontier AI models with 0 moat that are basically as good as Google's, a company which has the luxury of other revenue "streams" (rivers/seas/oceans) and can certainly win any price war with Anthropic.

    Can't wait to buy Puts on ANT...

  • xiaoyu2006 8 hours ago

    When you think the naming scheme is already bad enough... It can always get worse.

  • islemaster 5 hours ago

    As someone who works with real neon signs, the "Gemini 4 Neon" image at the top of the post is so unrealistic and jarring. It reads to me like a seven-fingered hand. Probably won't bother most people but it was a funny first impression.

    • vintermann 3 minutes ago

      That's got to be a niche market in this LED age. Or is it a hobby? Either way, I'm kind of envious. Real neon lights are cool.

  • bloky 8 hours ago

    At this point you just use roulette on which Gemini model you'll use for a task

  • simonw 6 hours ago

    Pelicans do indeed sit between Hydrogen and Carbon/Argon: https://tools.simonwillison.net/markdown-svg-renderer?url=ht...

    Here are the others in the family: https://static.simonwillison.net/static/2026/gemini-4-pelica...

  • firefoxd 4 hours ago

    Google Salesperson: ok adding Neon to the list... So you want to buy AI for your company [0]?

    [0]: https://idiallo.com/byte-size/buying-ai-for-your-company

  • WarmWash 5 hours ago

    I'm pretty sure Google is intentionally using non-descriptive names to get people to use the model they think is best for their task, because at this point, it probably is.

    On Google's scale you could probably mint a datacenter of free compute if you could get people to stop using 10T parameter giants to vibe check their emails.

  • ncd 8 hours ago

    Okay, but now will you let my tiny team sign up somewhere to be able to use this the way I use `claude` or `codex`? I'm told to use `antigravity`, but when I go to https://one.google.com/ai (which is where the magic answer from Google search told me to go), I'm told that I have to use my "personal" account, not my work account. So I clicked "Manage Google Workspace Account" on that page, and got dumped in a super dense UI with nothing related to enabling this model somewhere.

    Good lord, this is an insane flow.

    • smallerize 8 hours ago

      So I haven't tried this, but it looks like you have to set up the Gemini API in a Google Cloud project, then point your Antigravity at the project. https://antigravity.google/docs/pay-as-you-go/

      • ncd 7 hours ago

        Thanks for finding that. Looks like it would use API pricing? I was hoping to be able to get a subscription like offered by the other competing models, but might not be able to do that.

  • anuptalwalkar 3 hours ago

    I can’t wait till Google or a model in google invents a new element when they run out of the periodic table.

  • hiddenvulkcan 8 hours ago

    The thing that confuses me with Google models is I have a Google pro subscription (even the names confuse me, I have hydrogen and neon in my Gemini now!). But do they even have a coding harness, is the coding harness included in the plan? Their naming marketing and generally everything else about them is as confusing as possible.

    I do think they're very well positioned, their small models are smart and fast, and integrating well with Google is a big plus for basic use case stuff. I always like these models!

    • hhhAndrew 4 hours ago

      The coding harness is Antigravity or agy for the CLI + antigravity.google.com for remote control. Disclaimer: work there

    • Ternari 4 hours ago

      Antigravity. Yes, you can use it with your Pro plan.

  • xnx 8 hours ago

    It won't be good for Anthropic's stock price if investors realize it's always been a three-way race and Google is on a trajectory to overtake.

  • comboy 8 hours ago

    Gemini Chooser. Our frontier intelligence which ranks #2 in benchmark on choosing the right Google model for your task.

  • hendry 8 hours ago

    I only have 4 hydrogen and 3.1 pro in the app as a Google plus subscriber ~8gbp pcm.

    Nevermind the 20GBP pcm I pay for YouTube.

  • origini 8 hours ago

    Everyone uses something you can remember somehow than there is google. Not only do you hear their model names rarely (neon, argon etc. is new right?) but then they choose something which is not obvious at all.

    And then you have big names like Hy-dro-gen haija

  • thisisauserid 8 hours ago

    Good to see they are still releasing these... periodically!

  • hn8726 8 hours ago

    I'm very excited for any model that performs on par with Opus 5.5 but doesn't speak claudish. It increasingly looks like Google might get there soon

    • Ternari 4 hours ago

      They’re already there. Have you not used Carbon or Neon yet?

  • ryanschneider 5 hours ago

    I think GDM has clearly achieved ~RSI~ RSS (recursive self-sabotage).

  • jatindavis5 3 hours ago

    what are these new names being introduced all the time. neon, hydrogen, carbon etc.

  • clutter55561 8 hours ago

    Surely the AI companies are colluding... to make sure no one can make an informed about model class/size/effort.

  • avazhi an hour ago

    Go away Google, you're drunk.

    Maybe focus on 1 or 2 things at a time and do it right, eh?

  • wangii 5 hours ago

    wow, the top comment is a living example of bikeshedding!

    I am confused with the naming scheme as well, and it's a fun topic, right?

  • Jorge1o1 8 hours ago

    This is some of the most confusing naming I've ever seen. There's now: Gemini 4 Hydrogen, Carbon, Argon, and Neon!

    Maybe the atomic number correlates to model pricing? Nope! Neon (10) is marketed as cheaper than Carbon (14).

    Maybe the atomic number correlates to overall intelligence? Nope! Neon (10) and Carbon (14) are both marketed as smarter than Argon (18).

    The only thing I can think of right now is that it's in alphabetical order by release... but in that case why the periodic table?

  • freecodeio 8 hours ago

    something tells me this janky naming scheme will mean the whole gemini section of google is just gonna get shut down and replaced with something new during 2027

  • 45sdasf45 8 hours ago

    to paraphrase xkcd 2501: "Its easy to forget the average person only knows the atomic numbers of the noble gases and one or two nonmetallic elements".

    Say what you like about Apples Pro, Max and Pro Max -- at least it gives some indication of size to the average person.

  • cesarvarela 7 hours ago

    So basically:

    haiku -> hydrogen

    opus -> neon

    fable -> carbon

    No mention of Argon in that page though.

  • Revanche1367 7 hours ago

    The names just keep getting more random. And the table at the bottom, I'm pretty sure it's conventional (at least in the English speaking part of the world which I'm not originally from myself) to put your plan/model/price/whatever tiers left to right in increasing rate order. They do right to left here. Are they deliberately trying to confuse people? I get it that we tech workers are generally wary of google, microsoft, etc. naming conventions often (always) being terrible, but the average user out there (which increasingly is everyone with a smartphone or PC) could just skim through the announcement and go away with an easily avoidable misunderstanding.

  • nkmnz 8 hours ago

    Waiting for Gemini Plutonium to launch.

  • InsideOutSanta 8 hours ago

    Are these names a practical joke Google is playing on us, or are they actually serious?

    • maplet 8 hours ago

      My guess is that Google thinks the trade-offs are too nuanced to have a simple small/medium/large ordering. Neon is between Hydrogen and Carbon in terms of speed and cost, but it scores highest on Artificial Analysis Index and has a xhigh thinking level (unlike Carbon).

  • sidibe 3 hours ago

    You can count on Google releases now to have way more discussion of the names than the models

  • kristjansson 8 hours ago

    So now we've got to get a periodic table out to understand Gemini releases? Hydrogen is the weakest, and Argon is the middle, but Carbon is super-Argon and Neon is better than both? And I can't even look at the price list if my logged-in google account doesn't have AI Studio(??) access?

    e: No, Neon is the cheaper, mid-tier between Hydrogen and Argon. So Argon/Neon are aligned along the noble gasses, and Hydrogen kinda makes sense as the weakest ... but they threw Carbon in without any consideration

    • Melatonic 8 hours ago

      Maybe Carbon is actually just two Pelicans in a trenchcoat

  • dekhn 8 hours ago

    Waiting for Gemini 4 Oganesson

  • rednafi 8 hours ago

    And naming things remains one of the hardest problems in CS. At this point, I have no idea what to expect when I hear the names.

    Just give them a version number and an identifier, like:

    1. V5 smol 2. V5 mid 3. V5 big 4. V5 yuge

    WTF are Hydrogen, Oxygen, and Nitrogen? At least OpenAI and Ant have Luna, Terra, Sol, Astra and Haiku, Sonnet, Opus, Fable. They use those consistently.

  • lousken 7 hours ago

    so is this carbon but fixed or..?

  • moecables 8 hours ago

    It's crap how Google doesn't show pricing on the pricing page for Antigravity: https://antigravity.google/pricing

  • sashank_1509 7 hours ago

    Probably the worst naming scheme,

    Luna->Terra->Sol->Astra is fine, you need to be familiar with meaning of the names but it makes sense and works.

    Haiku->Sonnet->Opus->Fable/Mythos is better, more common words, more obvious understanding.

    Hydrogen->Argon->Neon->Carbon (is there a Barium too somewhere? ) fck all I have no idea, people don’t even know noble gases well, but hydrogen and carbon doesn’t even fit that. 1,2,3,4 would be a better naming scheme.

  • inshard 8 hours ago

    I don’t know what’s going on with those GPT 6.2 prices. There should be some standardization auditable in the x-axis cost per task. If not, labs may game that through loss leadership tactics.

  • TacticalCoder 8 hours ago

    An army of Anthropic shills is going to whine "benchmaxxing".

    That sounds like serious benchmaxxing for Gemini 4 Neon is ahead on 17 out of 19 benchmarks vs Opus 5.5.

    Not so bad for a company that, according to HN, "has lost the AI race".

    P.S: I like the Google models but a truce before the Anthropic IPO would have been nice. This is just a very big dick move a few hours ahead of Anthropic's IPO. Shall we see Anthropic launch a new model from now 'till the opening bell?

    • 875323897533 4 hours ago

      Stunning and brave of you to defend this small startup that will disrupt the AI ecosystem.

  • BrokenCogs 7 hours ago

    What's the point of Carbon now? Neon seems better than it in both cost and performance

  • Robdel12 8 hours ago

    Their pricing page has no.. pricing? Am I the only one that sees that?

  • hippycruncher22 8 hours ago

    Can someone explain wtf is going on at google?

    • mpvjs 8 hours ago

      gemini tungsten can

  • hackersnooze1 3 hours ago

    If you trust Google with anything you are a fool. Currently if you use the AI tab in google search, there is no button to submit your query and the enter key just creates a new line. You literally cannot submit a query, at all. Any company that is this fundamentally incompetent shouldn't be trusted with a penny.

  • lawandjustice 8 hours ago

    My predictions:

    2027 january RSI is achieved. 2027 february first big scary incident (100 million people severely affected by infrastucture collapse for several days) 2027 march a fleet of rogue self replicating agents is discovered. 2027 april first severe fatal incident (1000+ people dead, most likely from a missile created using oss models). 2027 june. Deadly virus created by a terrorist organization is released in Ukraine. Remind me!

  • newhotelowner 5 hours ago

    Using claude, we were able to learn how two of the software works, and build our own with API that can connect to our PMS.

    Gemini is not very helpful there. It just kind of gave up.