NGL, I think this might be my favorite benchmark, ha. While the Chinese labs asses are a bit off, I certainly have big hopes for their next versions. Their hearts are certainly in the right place.
What I'd kill for would be an app where you take a pic of scribbles and drawings on a whiteboard/notebook and the app turns it into SVGs you can then continue to manipulate digitally on PC/tablet.
They can represent something in vector, but it's always very hard to use the result for anything else. Like, the paths doesn't make sense, aren't grouped to the logic elements etc.
Last time I tried, it did not. That's why there's a place on the market for an app just tuned to handle the transition seamlessly, rather than having you fucking around with prompts in LLMs.
Beyond SVG, these things have been RL'd on 3d reconstruction (which carries over), and I expect given the public reaction to recent frontier advancements in 3d capabilities that this will get allocated a larger share of the post training budget in the future.
> How many pelican riding bicycle SVGs were there before this test existed? What if the training data is being polluted with all these wonky results...
But Simon has demonstrated before that asking for any animal doing something silly will still get similar results from the same model.
It's still clearly not a part of any model's training set in the sense of pelicanmaxxing in particular. It's either a capability to generate arbitrary SVGs (or even something broader), or no specific training at all.
Sadly, this is all that managers know to do: take a metric and try to optimize the heck out of it. They don't care about research, technology, "don't be evil", etc. And Google is run by managers now.
Expect Google models to be benchmaxxed as much as possible, but don't expect commensurate performance on real-world tasks.
Certainly true but I think the question is “does targeting the SVG pelican mean that the quality of all SVG output improves?” If so it’s a worthwhile thing to target.
It's a little bit like any benchmark, especially microbenchmarks. It is useful as far as it reflects the broader use cases, but the risk of overfitting to the specific task and not the broader ones is very high.
Recently, I’ve discovered that Claude can render great SVG versions of user interface screenshots. I ask it to substitute simple SVG friendly landscapes for any photographic images that happen to be part of the screenshots. The results are much better than they should be, especially with Opus. Usually good enough to be used on a website as a way to get an infinitely scalable, much smaller version of a screenshot. But clearly, I should try the Gemini models on this real world task. That’s where we’ll find out if all of this training to beat the benchmark actually pays off.
How many cicles before pelicans become photoreallistic SVGs and them come back to stylistic artistic abstract renderings?
Also, that first tweet contains a commentary that, for me, explain a lot about the public misconception about how current AI works:
> wait, does this mean you guys are actively training svg generation? i was mostly amazed by it thinking the models doing better at it by itself due to becoming better at reasoning.
No, it's not better reasoning that drives current AI, it's better examples we have been giving them. Whole army of people generating content for the frontier labs.
Also, reasoning improving the out of distribution part, like computer use, tool use etc.
As a cyclist, I tend to pay more attention to the bike shape than most people. Everyone knows what a bike looks like, but as with LLMs, most people would struggle to draw one - correctly connecting all the different tubes at correct angles etc.
I note that even these bikes are not actually rideable. For a bike to be stable, the front wheel axle needs to extend forward a little ahead of the fork - the fork must ”bend forwards”. This is what actually keeps the bike upright, not gyroscopic effects or some such. Given LLMs are generally good at basic maths, im surprised they get this wrong.
Maybe a bit less so in this case though. As per the announcements, the model was actually trained for this prompt, so was fed some training data. I can well imagine the data there was just not so good.
Is it me or is there maybe some carryover happening where the pelicans are portrayed as expending different degrees of effort too? Like, the high effort pelican has a determined look on his face.
Is the prompt context telling the models what their effort level is, and the ambient presence of that concept of ‘high effort’ in the context window is leaking through and affecting the portrayal?
xhigh: "Reasoning effort is set to xhigh. Please think carefully through the task, validate key assumptions, consider plausible alternatives, and prioritize correctness, consistency, and clarity in the final answer."
low: "Reasoning effort is set to low. Keep your thinking brief and focused, moving directly to the conclusion without unnecessary elaboration."
At some point Simon needs to move from Pelicans to something more complex. Every time a new model is launched by major Frontier company, I already know with certainty that first HN comment on the news would be about how Model is drawing the Pelicans.
Front bicycle wheel in all Argon examples is still wrong. This bicycle will not be stable. Axis of rotation of front wheel needs to be in-front of axis of rotation of steering wheel. Frontier VLMs still do not understand how bicycles work.
Why are all these AI's making the pelican's pouch so large though? Unless they have some huge fish in there or something they don't tend to look like that... Is it that it's drawing a "cartoon" pelican or something?
It's obvious OAI hillclimbs the image generation -> SVG loop. I'd put money down on that being why OAI's pelicans have a yellowish tint that matches the yellow tint of OAI's earlier image models.
simonw, your pelicans are so famous now that the frontier labs can just RLVR for SVG images explicitly.
It's a simple manner of prompt...generate concept art with nanobanana, measure, llm creates svg with text, measure the distance between rasterized svg embedding to nanobanana art....do post-training RL on a 3 million such images and matching embeddings.
I swear this will be a canonical example of Goodhardt's Law in action.
You have to switch to a new benchmark that isn't so easy to run a verifiable rewards pipeline on...Perhaps a longer horizon task where it's easy to mess up midway.
Why? Don’t you think the steady improvement in pelican bicycling SVG art is itself a worthy field of endeavor for LLMs to be optimized for?
Yes, Simon has made pelicans a benchmark and now models have grown very good at rendering them. Is there a reason to believe that this isn’t manifesting as useful skill in general in SVG generation?
It starts to be so good that the model doesn't really matter. Each model starts to have its own personality, and it's the personality of the drawing, like a real artist, that defines the style of the pelican, and that makes it harder and harder to decide if it's good or bad.
The models that are really good at 3D modeling tend to also be really good at pelicans. I think the skills are strongly related - it's about having good 2D (and now 3D) spatial awareness.
The test is always going to be hard though, because it's an absurd image to ask for.
"Pelican on bicycle in SVG" is an adversarial example on multiple levels - out of distribution, ill choose of tool for the task, and anatomical impossibility of real pelicans using typical bycicles.
Yeah, that's a good point. My basic pelican test doesn't allow the model to render their work and iterate on it. The 3D models all come from agent systems that are rendering with Blender or screenshotting with Playwright along the way.
I’d really like to see what each model comes up with if given a basic harness to render and iterate. They’ve all gotten pretty good at basic one shot pelicans, but seeing what they do when given a direct goal, and even getting a time lapse of each image as they edit and adjust would be really fun.
A lot depends on what the models were specifically trained for using reinforcement learning. Modern models are generally OK and everything, but they're only really good at the things they have extensive RL for.
Honestly surprising if other labs aren’t training for svg generation. Feel like SVG’s are going to be really useful for dynamic UI generation stuff that goes beyond cobbling Html together
Buddy, there's a Pelican riding it. No real cloud is like that, no real sea is like that, no real gulls are like that. And that's fine because it's a cartoon.
what is even the point of these now? labs are most definitely training on this question now. I feel like this test scratches the surface of a "real" test for these new models
It attracts at least one comment like yours on every single thread, so I guess the main purpose is to kick off a race to see who can post that comment first.
People who are unable to comment on the underlying technology or capabilities because they don't know or don't care to inform themselves sure can feel good that they have an opinion on pelican art.
The roof is always raising month over month. If you were to make a harder version of the benchmark, what do you think you'd like to see the agents do better?
Medium and high are fixies apparently, (or they have a worm gear, although the hub doesn't look big enough for that), since neither has a derailleur. That would also explain the lack of brakes at least.
Low is still obviously failing the "test". It's pedals are not actually attached to the bike either.
Medium has a handlebar that goes through the pelican's beak. So I'd say also failing.
For the high, the only nitpick would be that the bell isn't quite on the handlebars and the stripes on his scarf don't line up correctly.
The models are solving Millennium Prize questions and making progress on the Riemann Hypothesis, yet the top comment is still looking at pelicans somehow. To me, it feels even more superficial and pointless than arguing about the presidential candidates' tie choices in a debate.
> The model launches today at an introductory price of $2 per million input tokens and $10 per million output tokens (with cached input tokens at 95% off).
Couldn't you just say what the price will be after the introductory period and when it'll take effect? Do they have some SKILL.md for these blog posts that requires the most important information to always be buried ten levels deep and hidden behind a login page?
With the pace they're delivering models, there's gonna be 1-2 new Gemini 4 versions by December 1st, presumably with their own introductory price windows. We just got Gemini 3.8 on 2 September and Gemini 3.7 on 13 August.
This reminds me of telephone commercials that ran back around 1987, saying they were offering long distance calls for 10 cents a minute (a pretty good price back then) and -- this was emphasized, a huge selling point -- they guaranteed your rate would never go up. In retrospect, all the lols :-)
I wonder if that’s more to price in the cost of keeping old models around on expensive inference hardware. It’s not like old models get cheaper to run, and a more heterogeneous fleet is less efficient to manage.
> Couldn't you just say what the price will be after the introductory period
+1. FWIW, I didn't even get far enough to wonder about the price.
The lead (and only) sentence states "Our next era of frontier intelligence, built to solve your hardest problems" a statement equally true of any new Google or competitor model, hence completely devoid of useful meaning. I then kept scrolling down... and down... and down, giving up after a half dozen screens of big images, huge videos and some customer stories. The first sentence of every text block was equally vague or meaningless so... tv;dr (Too vague, didn't read).
I'm using this model via API. I am most excited about its ability to write the english language (or any language). My least favorite trend on twitter has been people posting 2D illustrations/3D renders as a proxy for model capability. Well-written prose is far less likely to go viral but far outstrips the importance rendering a fantasy world in one shot.
I've been trying it out via the API and the model has been a delight so far. Clear in the head, not RL-fried, distinct writing style, and big model smell. That said, the em-dash is back w/ this model.
That said, I would be so frustrated with Sundar Pichai if I was on the Google team. The botched rollout of this model seems to be entirely a matter of capacity, and Pichai has been more than happy to give it to anybody but his own team.
I've been using it for a while internally at Google. I prefer Argon over Opus 5.5 because of the clear thinking and directness. It phrases things in an organized, straightforward way.
Announced the model two weeks ago to a ton of hype, model was unavailable for anyone to try. On twitter, Google folks went on a great run taking shots at Muse ("Muse who?"). The hype died down bc nobody could touch the models and meta/muse came out with a similarly capable model in the meanwhile.
At the same time, OpenAI had a disastrous dev day that led a lot of users to jump ship to Claude.
Couple weeks later, Gemini-4-argon is still unavailable on anything less than the ultra tier, which means there's no natural funnel to get ppl to pay for the more expensive ultra plan other than "trust us the model is good".
If Google had argon broadly available earlier this month they would have picked up a ton of new users.
As someone internal at Google I don't understand why there's so much criticism of Google models... but right now I'm using the next model beyond Argon, and Flash 3.8. In my personal Gemini app I only have access to 3.5.
Models move so fast now and Google is very cautious about releasing them.
All the power users are always jumping around by now and many hold the smallest subscription with all the models or at least two.
Narrating this as if it’s the game of thrones is a fun exercise, but it’s just becoming commonplace now that models jump ahead of each other on a weekly to Monthly basis.
> and many hold the smallest subscription with all the models or at least two
This was part of the issue, as stated in their comment. We couldn't jump around to try it for the reason that is wasn't made available to, as you say, "all the power users".
Gemini 4 is announced be available to Pro users in the Gemini app.
I mean, I heard rumors too that it would be ultra only, but that doesn't seem to be the case, unless you are specifically talking about antigravity?
Realistically, market adoption isn't going to be the result of terminally online people (like me, I confess) following things way too closely on social.
There weren't "rumors" that it would be Ultra only, that's what Google has announced two weeks ago. Happy to see they've went back on that, but I haven't seen Gemini 4 on any apps till now, and I'm a Google AI Pro subscriber.
In particular, where cost is the x axis, a well-funded company with a good but not best model can discount their new model until it reaches the frontier. In some sense this is as much a statement about the efficiency of markets in the presence of public measurement-based comparison and public list pricing as it is about research pace.
I just created a Google account, which was a hassle. I tried to find an app and found AI Studio, but I’m not sure what that is. Then I found Antigravity, which apparently is Google’s alternative to Codex, and I installed it. Then I tried to buy the small Gemini AI Ultra package, but it didn’t let me pay. It accepted neither my credit card nor my PayPal account. It’s crazy how bad the Google experience is nowadays. You don’t know what to use or where to find it, and the security measures seem totally overkill. Am I really the only one having these experiences? Is Google aware of that?
Same thing, I've always paid my bills, I have had some sort of google subscription for the last 10 years. Now they're requesting me to switch to pre-paid billing. They won't accept my card (that has worked every time this year so far).
Is there anybody in Google having a look at how bad and complicated their payment system is? In all the other AI providers you just sign in, pay, get a key, done. In other clouds like AWS, you add a card, pay little to verify the card the first time, monthly bill, done. In Google it looks like you have to fight them every time for the very thing they should make as easy as possible: opening your wallet.
Using google services has always been miserable. It used to be way worse. Now you can at least get an api key from AI studio without going through 20 esoteric steps to configure everything.
I hate everything with Google and consumer facing services, they dont give a shit because 99.9% are monetized through spyware or are buisness customers.
They had gemini cli, what happened to that? Was fine
Interesting! Do you have webgl disabled for incognito or for chrome in general? I wonder if it's trying to do software rendering.
FWIW, I have literally the exact same configuration (M3 air, google chrome). Here's a recording on mine where I'm seeing 60fps: https://non.io/video/video-3587018401.mp4
Truly S-tier. It's sharp, creative, intuitive. Can't believe how much better it is than Opus 5, then they went ahead and made it cheaper and faster too.
I'm constantly blown away by it. I do think Astra is slightly more intelligent (although worse at this particular benchmark), but Opus 5.5 has become my default model for 99% of what I do. It's so good.
It's a great model. I have had it trying to decrypt the Voynich manuscript for the past couple of days without being biased by other research/other people's attempts. If anyone is interested: https://github.com/d2kx/voynich-research
> An interactive companion to our ongoing exploration
Eyes glazed over again.
Did you even read your own site before you posted it?
One of the first pages I opened:
> Smoking-gun proof of plant-label ingestion: in Row 2, the central cut root is labeled okol (f88r.15), and directly underneath it on Line 2 (f88r.19) the recipe writes...chokol daiin qoekol qoekol qockhol okol cheol...!
How about you post something you've done/built instead of just insulting other people. Spend your mornings using people stuff just to poke holes in it?
The "Explore the latest" section is nearly blank for me - are there supposed to be diagrams or visual aids, or is it one headline and sentence taking up the whole page on cards?
Gemini isn't perfect but I find it a lot less annoying than other model families (personal preference or stockholm syndrome). Interested to read the reviews.
Seconded. Anthropic feel the need to limit how abusive you can be to Claude because god, I want to throttle that snotty little shit. Google don't have to do that with Gemini because it just isn't as irritating. It does what I ask and doesn't try to second-guess my requests, or do "the quick fix".
It's been really bad at lyrics but gotten better. You look something up and it will just make something up and it takes a bit of digging to discover it's because of copyright
Good God, man. Google has the worst frontier lab rollout of them all. And they're delivering products that are behind the curve constantly.
Now, I happen to believe that's an excellent strategy. I bet they actually resolve more LLM requests than any other company at this point. The lay person still uses Google exclusively.
just cancelled AI pro myself, since I exclusively used it for AI and also pay for youtube premium family it was becoming quite the monthly just for google
if they throw in youtube premium in that package then I might reconisder
Google One - Pro tier. I have it via Lenny's Product Pass which, imo, is a much better deal since it includes Cursor Pro, Granola, and like dozens of other paid sub tiers: https://www.lennysproductpass.com/
I use the hell out of both. Kind of sad that they took Spark out of beta-access, being able to orchestrate via your phone on something that has direct access to your Drive meant I was authoring skills and other functionality on a schedule with spreadsheets serving as databases, deterministic tests, etc.
Remembering the horror stories about lost Google accounts, I would be comfortable paying the next lower tier (100 GB) because I can backup everything myself, but not the 5TB plan, I'm sure I'll lose something if the account is lost. Hopefully these plans guarantee a human in the tech support loop.
Yeah, very true. I need to set up a recurring backup-to-external-HD automation for it. It's vanishingly easy with the app and I don't doubt they're training on all the data itself.
Google AI pro in the USA, if you have Verizon wireless it’s $10 a month add on. Without Verizon it’s $20 a month I believe. You can also share most of the benefits with up to 5? People.
Afaict, they are targeting difference user bases, Google being after their existing user base, having lost in the developer ecosystem to the two labs and open weight models.
Corporate companies. Bog-standard office workers don't need top-of-the-line Mythos intelligence, just something to add to Google Workspace. I personally doubt developers use Gemini for coding beyond marketing materials.
Are you suggesting it’s BETTER than their competitors?
Because if it’s not, and it’s not available to all paying users (Right now, it’s only available to Ultra users), then they are behind the curve if they’re treating it like Fable.
Maybe? But it's so many orders of magnitude apart. I can easily do 1B tokens a day (output tokens is probably like 3M?).
When I see non technical people use that Google reply I'd be surprised if it's more than 1k output tokens, and with a tiny model. So a 3000x in output tokens, on a model at least 10x as expensive, I wouldn't be surprised if hardware usage is 30.000x for a dev that uses coding subs vs someone who uses the Google AI overview a couple of times a day.
And that's not counting the absurd difference in cache reads.
Moreover, you need a Google account in order to use it. I wouldn't bother paying 200 USD for it but I just don't want an entire google tracking package just for using one product.
Oh, I didn't notice that they switched from 1.5 to 3 flash about six months ago. I know someone there and they talked about how 1.5 was hardened and it was a pain to do so on newer models that is wasn't worth it for a long time. But that convo was in the beginning of the year. I should have looked it up.
Of all models I like to write with Gemini the most. The language appears much more natural and elegant than what even Opus 5.5 or Astra do. Also, they give 5TB of cloud storage and discounts, it’s a steal for what you get.
Agreed. It's much better at working with stuff (like documents and presentations) in other languages too. It feels like it's trained in multiple languages from the ground up rather than as an afterthought.
Adding my five cents: initial usage through OpenRouter has been great. Very fun model to spar with, clever, intelligent. On par with Opus 5.5 but without the Claudism's and more of a Sol type of conversation. Though speaking of Sol, I've been very dissapointed with Sol 6.1 to the extent that I feel OpenAI has neutered it compared to a few months ago. Especially when giving it goals it will happily run in dumb circles just burning tokens without any real work.
The model used is irrelevant when there's the critical issue where their "AI" assistant can't understand the word "nearest". Literally worse than a second-grade understanding of English.
I also have had shockingly bad results with gemini on android. Ended up giving up on it after a few months. Silly silly google with their trademark UX polish and attention to usability.
This reflects pretty much every result of mine from Google these days. Not just navigation, but search too.
The only thing I can figure is that Google's consumer-facing free-to-use models are set to "reason as fast as possible with only the most basic plausibility checks." I have to correct them like this every single time.
this is like that tiktok guy who asks the spoken ChatGPT model very simple questions, like 'count down from 60', or 'how many e's in seventeen' or whatever and the model shits the bed every time it's hilarious
I think these companies have good models but expose them poorly like others have mentioned
Often it seems like the problem with the Gemini apps is the harness, not the model itself. For example if I upload an image PDF it recognizes text worse than if I use the API.
Is this from you car ? Gemini in car is probably less intelligent than a spider.
I tell it to "Play Let it go from frozen". It plays it on youtube music. 5 seconds later I tell it again 'Play let it go from frozen" it then just dies saying "I can't play videos in car".
The auto version is extra horrible because what it reads to you is just a summary of what the model actually said. I don't ask it anything when driving because it's too frustrating.
Yeah, whatever they have running on Google Maps is probably a pretty low tier model, but it is total trash. I’ve only used it twice though and that’s when it first came out. I don’t know if there have been any changes since then. The main problem is I felt like it was not keeping track of the context of the session.
which is more or less the crux of the problem android auto's gemini implementation. It sometimes does not even know that it is in a car. The ghome implementation is also like this.
To everyone saying not available yet: it looks from the page like Pro and Ultra subscribers have it today and given the play links, it'll probably show up in the API in a few hours.
Available in agy for me as of now, I'm running my new model benmark of having it do comprehensive code audit on one of my more complex projects, it spun up 1 subagent per audit section and it took ~10 minutes per subagent on an 800k LOC, 60k complexity, ~4k file project. Will take me much longer to go through and determine the quality of the audit work.
Antigravity CLI 1.3.4
[My account email] (Google AI Ultra)
Gemini 4 Argon (High)
~/[working folder]
> /model
⎿ Model set to Gemini 4 Argon (High)
Same. It just told me "agent execution terminated due to an error". There was a little chevron underneath it that let me expand the error to view more. I was hoping to see what the error was, so I clicked it, and it said... "agent execution terminated due to an error".
It's crazy to me how care there is in most of these vibe-coded vibe-coding harnesses. I don't like oai's models much but I have to say their UI has by far the best attention to detail of any major player.
I'm on Pro and it's trying to get me to upgrade to Ultra in the AI Studio UI when using Argon. In Antigravity CLI and the Gemini website it's not showing up at all despite being logged in
I like this phrasing and am curious what new word or phrase we might use to describe big companies over pushing Ai internally at the expense of their product quality. This pattern is so ubiquitous across google cloud products I have taken to calling them the 21st century IBM, having been a IBM baby and witnessing first hand their fall from exceptionalism.
The internet and gaming has ruined my brain because scuffed was a way to describe a bad dungeon run in mythic plus dungeons in WoW, or a bad tower dive in LoL a few years ago. I see scuffed and think "here's a WoW/League knower".
HP is the other company to consider as analogous. And maybe more explicitly so, as they had a similar "engineering first" internal culture that degraded markedly over time to be a shadow of their former self.
And took similar companies that were quality engineering focused -- like DEC or Palm -- down with them through acquisition.
I couldn't see it either, then I clicked the 'Antigravity' button in the top right, switched back to the main screen, and it appeared in the model selector.
"Argon autonomously turns raw video to a physics-accurate simulation, estimating depth and physical boundaries, and writing the exact simulation code required to replicate those real-world environments."
Geometry could be understood today with the model but what also really matters is the sensing, friction, mass how smooth the subjects are, dexterity is a very difficult problem so solve, not sure how simulations could handle such factors
Game engines have many of these features. At one point, iirc, Unity was positioning to be the simulation environment for world model agents before we spoke about it in these words.
I have the Free AI Pro plan from when I purchased my Pixel 11 Pro. I still don't have Argon in AGY nor web nor Gemini app. I have a free product and still feel cheated.
I'm open to Google winning me over again, but there isn't enough hype in the world to do it. They will need to earn it after these disappointing 6 months.
I appreciate how it's being integrated into the Google ecosystem, but the user experience still feels unpolished when actually using the system. The marketing, however, is impressive.
can for the love of god we get some usage in cheaper subscriptions being Ultra tier only is annoying. I want to just test it out here, well will burn some money via api I guess.
When you are on "gemini.google.com" there is a little lock icon next to Gemini 3 and Gemini 4. When you click it you get the prompt to subscribe (pay). Once you have successfully done so, the Gemini 4 option disappears!
I seriously doubt their ability to double API prices on Dec 1. Token prices are in a clear downward trend and Google should maximize revenue by competing for volume. Currently OpenAI is the only US frontier lab trying to seriously compete with open-weight models that you can run yourself for an order of magnitude lower than what frontier labs want to charge
Is there a benchmark for reverse engineering compiled programs?
- Grader compiles a program to binary -> Agent has to reverse it back, and match all bugs and behavior.
Because you have an android phone and don’t want to mess around with anything else(i.e the same reason people just use internet explorer on windows) or you hold Google stock so you want to support your investment :)
Really? Reminds me of Gemini 3. It took so long before being able to get a subscription for it, and then I would run out of their equivalent of the 5-hour usage window in a heartbeat.
Yeah maybe I've just become used to how much I can use Flash 3.7/3.8 but on the less expensive Ultra plan I basically never exhaust my quota even when working over multiple projects, but this was pretty much immediate. I did ask it to audit a project but still, I do that in Flash regularly and it's not especially costly.
I gave the model my usual question of "I got a raise from X to Y, please explain what is the impact on my net salary". Answer was detailed and nailed all payroll edge cases, only to miserably fail in basic math at the end. It tried to calculate 430 minus 60 and got 410.
Only available on Ultra, so I thought maybe I'll try it for one month to see if it's good. Well, I have Google One and for some reason this means I cannot sign up for AI Ultra? This is just one of the worst product rollouts I've ever seen
I checked gemini, set it to "gemini 4 argon" and it offered me Pro and Ultra to get access. I checked the normal plans, saw there is also plus.
Okay, so I'll get Pro to try it.
I subscribe.
No gemini 4.
The most generous reading I have is that it's a mistake. But I also downloaded the PC app, and now I really wonder why I did that, it seems to offer nothing beyond the web experience and also doesn't even know its a local app. (When asked about its capabilities it always responded with reasons pertaining to it being a web app.)
Right now there are two top posts one about GLM-Legal and one about Argon. Both score about the same in the Harvey Lab test, but the prices are $0.15 versus $10.55, the $0.15 is Vals AI's estimate using the base model since GLM-Legal has not published prices.
Yeah, but one is a general purpose model, and the other one is specifically geared towards the legal field. Considering Opus 5.5 only gets a 4% on the test compared to Gemini’s 20%, that’s actually pretty incredible.
I would love to try it out, and the Gemini models had previously been some of the ones I'd liked best in an overall sense, but unfortunately their updated inference endpoint is no longer ZDR on OpenRouter (change takes effect on the 18th). I think this is one of the more unfortunate turns in AI development in 2026, as it seemed like labs offering it was becoming the norm last year, and Google had previously done so consistently.
I hope all the people very concerned about AI safety are cheering this one. The best guard against one AI swarm taking over is having multiple adversarial orgs all with similarly advanced AI swarms.
I wonder if any of the AI safety people will begin encouraging AI researchers to swap jobs to work at a failing lab in order to keep them all extremely competitive with one another.
Digging in the model card, "Gemini 4 is the next generation in the Gemini series, a suite of highly-capable, natively multimodal, reasoning models. Gemini 4 Argon delivers frontier performance..." If they're using noble gases there are 2 obvious lighter naming candidates and 3 larger (or 4 if you count Oganesson). I'm not seeing them claiming Argon is the top/biggest/bestest model among the Gemini 4 series, just that it has frontier performance.
Edit: noble gases not noble taxes, though the latter would be a funnier naming scheme.
I'm still waiting for a model which I can tell to "take these 3 screenshots and make a product launch video", where it doesn't randomly modify the screenshots or garble/rewrite the text as it moves the screenshots around the video.
Every time I try to generate one it's like a slot machine, where a different part will be messed up after fixing the first.
I dont have that problem with opus5.5. I dont even take the screenshots anymore (it does all screen recordings)... I tell it a simple narrative and it follows through. e.g. https://x.com/plannotator/status/2108589088139886769?s=20
> Across Google and our early access partners, Argon is quickly becoming the preferred model for engineers tackling long-horizon software engineering.
"Across Google", wonder how much that is because of an effort by Google to curb usage of Anthropic models. MSFT and Meta similarly are pushing for in-house models. It doesn't mean the models are better when you push your engineers to dogfood and restrict access to other LLMs.
(Not that I know for sure this is the case, it's more of a hypothesis).
It passed the test of using latest crystal-lang concurrent feature (with web search) when gemini 3.8 flash failed (again, with web search enabled). I think it is at least useful for coding, for sometime.
Pricing is interesting. At launch Argon is the same price as GPT-6 Sol, half the price of Claude Opus 5.5 and 1/5th the price of Claude Fable 5.1 and GPT-6 Astra.
But on December 1st Argon doubles in price - so it will be the same price as Opus 5.5.
I've kinda stopped looking at token pricing. Argon might be the same token pricing as Sol, but in Artificial Analysis' testing, Argon costs 5.8x more to run their coding tasks and 2.2x more for their general index.
Likewise, you look at Artificial Analysis' general index and it cost $8,708 to run with Opus and $2,407 to run with Argon so even if Argon's token price doubles, it won't be the same price as Opus.
Sonnet is cheaper than Opus per-token, but in their coding tests Sonnet burns through enough tokens that it's 7.5% more expensive.
So much depends on how token efficient something is and whether it gets you a good result or not, right?
Plus, if one is using a plan instead of paying token rates, what does my $200 get me from OpenAI, Anthropic, or Google?
It just feels very hard to compare pricing given that the number of tokens used can vary so much. In Artificial Analysis' coding tests, Sonnet used 7.1x more tokens than Sol. Sonnet and Sol might have the same per-token pricing, but there's a huge difference in the number of tokens used.
All AI pages now include canvas animations. There isn't a single page without one. This is neither good nor bad, just an observation. I think OpenAI has the best animations so far, with Anthropic in second place. This one looks a bit off.
It seems the model is a little verbose, using more tokens to complete a task that other similar models as stated more precisely in the linked article about Artificial Analysis's work on this model.
Page says it is now available to Ultra and Pro subscribers. Although in AI Studio it currently says you need Ultra. But it's technically released either way.
Yeah, I think this is it! Google tried to hype their model but they have no chance against queen Dario. Perhaps they should just delete Gmail and say that Gemini did it. That way they would get some attention …
The slides they show in the catering video demo are so bad (in addition to having all the usual signs of AI design), I would genuinely laugh if a supplier sent me this.
Finally, gemini 3.8 has been falling way behind opus and sol in my experience. It's showing up in my antigravity. My work account though still has 3.6, 3.8, 3.1.
The previous gen pro used to have this weird tick where if you ask it a reasonably open ended question, it would number things in the response and assign each one some weird label starting with 'The...'
e.g.
X: The Parameters
Y: The "System"
Z: The "Ongoing Grind"
The first two prompts I tested in AI Studio does the same, unfortunately.
Another quite unpleasant tick was that if you specify a peculiar term in your prompt as a conversational shortcut, it would keep repeating that throughout the conversation. e.g. "...What is a bitter lesson pilled way to do this?", then you'd see it will quote "bitter lesson pilled" at like 7 places in its response. This one seems to have eased off a little.
Just switched, and it seems to be much better. I'm hoping it stops lying. I have Pro max subs for Claude, Codex, Grok, and Google AI... I was forced into using Google by running out of credits elsewhere.
I have found that Gemini often claims it did something, but didn't actually complete the task.
I just had 3.8 Flash do a major refactor -- that never should have needed to be done in the first place -- and switched to Argon 4, and had it review it.
Almost everything had been stubbed out by 3.8, and none of it was properly integrated into the codebase. :facepalm:
Glad to see them abandon the absolute train wreck of their model naming "strategy".
Gemini Live vs. Gemini Lite vs. Gemini Pro vs. Gemini Flash vs. Nano Banana Lite (that one apparently has no Gemini prefix?)... Please, just pick a metaphor and stick with it without wildly mixing reasoning modes, model sizes, and input/output modalities in the model name!
Unlike most people it seems, I like all the progress in AI but I really do wish they would reign in the over-the-top marketing. I mean, can we just get this stuff to the point where I can reliably work with it to design a well-architected software project without having to constantly struggle against bloat and nonsense directions no experienced or sane developer would ever take.
fyi: argon is an interesting choice of name, coming from the Greek argos meaning "lazy or living w/out labor".
>"Argon helped the Chrome team separate core webpage handling from a ~6 million line legacy folder of browser features held within Chromium’s 40 million-line codebase. Piping multiple clean code changes at a time, Argon helped the team finish an architectural refactor across hundreds of files, completing months of work in weeks."
Now that needed to be done for some time!
Regardless of if an AI helped or not, regardless of whether that was Google's or some other AI or not (though Argon does sound pretty cool!), regardless of whether an AI cut the time from months to weeks or not, refactoring Chrome -- decomposing it into individual clearly demarcated non-tightly coupled software parts with core webpage handling at the center -- that needed to be done for some time!
Now all of that being said, Gemini 4 Argon does sound pretty cool, and we congratulate both teams at Google -- the team that created Gemini 4 Argon, and the team that refactored Chrome with it!
But all of that being said, while the advancements made for Gemini 4 Argon are important and impressive (and they are!)... refactoring Chrome properly (IMHO) is slightly more important and impressive!
(You know, one man's humble opinion! That is, I am not the guy who is responsible for awarding performance-based bonuses at Google at the end of the year, so take everything I say with the proverbial "grain of salt"! :-) I am probably biased, one way or the other! :-) )
Why can't the Gemini model code?
I'd guess Gemini probably used the best dataset.
Even Google is the place where the world's best coders are gathered, isn't it?
Yet the GPT code from OpenAI, which collaborated with Microsoft, is beautiful, while the Gemini model's coding is trash.
I have no problem with Gemini Flash 3.8 for coding, though I only use Elixir which is one of the better languages for LLMs. Where does it fall short elsewhere?
Meanwhile I paid a subscription and they only provide access to 3.6 Flash. Consequently, I have moved on to open-weight models. It is what happens when you screw your customers over.
3.1 Gemini was amazing for anything non-agentic. I switched to it for my primary chatbot purposes.
I never touch 3.8, it's awful. Curious to try this, but I've yet to use any Gemini that is good for agentic work.
Also, unrelated but this made me look at Artificial Analysis to see where it lands on Cost per Task, and it ranks Opus 5.5 at almost double Astra 6. That does not align with my experience.
Not available for AI Pro subscribers in Antigravity. What a disappointment.
Google...why. You keep getting so close but constantly fumble at the last moment. I can use Opus 5.5 and Sol/Astra on the $20 Anthropic and OpenAI plans.
You are already playing from behind. You need to GET YOUR MODELS INTO MORE PEOPLES HANDS by any means possible. Continuing to make it difficult to use your products DOES NOT HELP.
This model might be great but the majority aren't going to subscribe for $100/month just to try it.
Blatantly obvious that models will not slow down. For any task, if we can build a dataset or environment of verifications, it can be automated. This will eat all of human tasks.
The fact that multiple companies can breach any benchmark means there is no ambiguity in terms of how to build AI, its just a matter of time.
Certainly Meta and Google are benchmaxxing to please public markets, but it still proves the above
but they have slowed down. the delta in capability between model releases now is smaller than ever before and i expect it to continue to diminish until we reach some kind of asymptotic maximum, which hopefully will mean turn towards optimization and local model improvements. The current frontier models running at 10x the tok/s would be incredible for my use cases
The delta isn't really slower, we're just unlocking capability bundles in batches. The capabilities bundles you care about might have mostly been unlocked already, and just need refinement, which is coming slower than the initial capabilities came, but don't discount the breadth of advances being made across all possible tasks.
We're not living in the same universe. Look at capability year-over-year. The asymptotic maximum is wishful thinking. Someone needs to step in before this gets wildly out of control.
> Someone needs to step in before this gets wildly out of control.
This sounds like a paid bot, lmao.
I wonder how much is actual improvements in the model and how much comes from better harnesses. I'd be willing to bet that the combination of better harnesses + the models being trained around agentic workflows (i.e the harnesses) has made it significantly better in many things, but it's hard to find other improvements imo.
You can see the difference between GPT 3, 4, 5, 6 each year. Same with Gemini and Claude. Look at image and text generation. Code generation. Math ability (Navier Stokes, Quasi-Riemann.)
Year-over-year, yes I would agree with you, but that is not the cycle time of models. And while I share the same hunch as GP, I'm waiting another cycle before putting more confidence behind it.
A little over a year ago they were winning their first gold in an elite high school math competition. Have you not seen what they are doing in math now?
i burn through a $200 claude and $200 codex plan every week. opus 5.5 is not "step change" it's just a bit better, and the benchmarks show this. Honestly with my workflow I cannot really tell the difference between the top models, they're all pretty much the same in capability, just a bit different. You can build software with all of them, and with all of them you have to validate the change thoroughly to not ship stupid bugs
I do not see the current training regime of "relentlessly try anything to get the task done" as a true step change in intelligence. This is one of the reasons I suspect capability acquisition is slowing.
A more intelligent model would not default to spaghetti throwing, it would pause to consider the problem fundamentals, the accuracy of assumptions, and ask questions that remove ambiguity. This however does not lend itself to increasing token sales while prices remain inflated from market and brand dominance.
Models just solved a Millennium price and you think their progress is slowing down? I will concede that their capabilities vary a lot from one field to another, but still.
Unsurprisingly, the pelicans are excellent. Google are the only lab who have publicly shared that they have specifically trained for SVG output - https://twitter.com/sunjiao123sun_/status/202455551655137292... - and they've had great pelican results in the past too (which they have shown off themselves - https://twitter.com/JeffDean/status/2024525132266688757).
Here's low, medium, high. Even low is better than pelicans from any other lab: https://tools.simonwillison.net/markdown-svg-renderer?url=ht...
low cost 10.185 cents, medium was 11.098 cents, high was 17.166 cents.
The costs are similar because it used more output tokens than reasoning tokens for all three pelicans, even for high.
Me in 2025 https://news.ycombinator.com/item?id=44759781
> If it's on HN and is a meme at this point, it will end up in the training set.
> It's kind of fun to imagine that there is an intern in every AI company furiously trying to get nice looking svg pelicans on bicycles.
Google listened and assigned a few interns, I guess, for the task
That's exactly what I was thinking when I saw https://www.assbench.com/ posted a while back: "oh no, this is going to be somebody's job".
prompt is to detailed (llm written?) better test of model creativity is to not constrain prompt
NGL, I think this might be my favorite benchmark, ha. While the Chinese labs asses are a bit off, I certainly have big hopes for their next versions. Their hearts are certainly in the right place.
omg that's hilarious
SVG is useful for actual tasks, not just looking good in a silly "benchmark"
Of course it would be nice if google actually added SVG support to google slides...
What I'd kill for would be an app where you take a pic of scribbles and drawings on a whiteboard/notebook and the app turns it into SVGs you can then continue to manipulate digitally on PC/tablet.
Inkscape and https://github.com/visioncortex/vtracer can both do this.
They can represent something in vector, but it's always very hard to use the result for anything else. Like, the paths doesn't make sense, aren't grouped to the logic elements etc.
Why would you kill for this? Why not just generate svgs with any editing software?
>Why would you kill for this?
Figure of speech.
>Why not just generate svgs with any editing software?
How do you do that from the whiteboard/notebook?
ChatGPT does that decently well
Last time I tried, it did not. That's why there's a place on the market for an app just tuned to handle the transition seamlessly, rather than having you fucking around with prompts in LLMs.
Beyond SVG, these things have been RL'd on 3d reconstruction (which carries over), and I expect given the public reaction to recent frontier advancements in 3d capabilities that this will get allocated a larger share of the post training budget in the future.
We all said it: https://news.ycombinator.com/item?id=46978726 (8 months ago)
> How many pelican riding bicycle SVGs were there before this test existed? What if the training data is being polluted with all these wonky results...
But Simon has demonstrated before that asking for any animal doing something silly will still get similar results from the same model.
It's still clearly not a part of any model's training set in the sense of pelicanmaxxing in particular. It's either a capability to generate arbitrary SVGs (or even something broader), or no specific training at all.
Yet, they're all almost exactly the same in style and composition.
Sadly, this is all that managers know to do: take a metric and try to optimize the heck out of it. They don't care about research, technology, "don't be evil", etc. And Google is run by managers now.
Expect Google models to be benchmaxxed as much as possible, but don't expect commensurate performance on real-world tasks.
>> this is all that managers know to do: take a metric and try to optimize the heck out of it. They don't care about research, technology
This is not true. It is needless negative hyperbole that IMHO doesn't belong on HN.
I'm speaking from 30 years experience in the Valley, having worked at several big companies.
Certainly true but I think the question is “does targeting the SVG pelican mean that the quality of all SVG output improves?” If so it’s a worthwhile thing to target.
It's a little bit like any benchmark, especially microbenchmarks. It is useful as far as it reflects the broader use cases, but the risk of overfitting to the specific task and not the broader ones is very high.
https://xkcd.com/2899/
They said they are training for svg not pelican svg specifically
Maybe so, and yet it's still not enough -- the pelicans all have flaws.
They look remarkably good, sure -- but not perfect.
I end up getting SVG output for all kinds of things, and never once have I asked for, or gotten, a pelican.
Recently, I’ve discovered that Claude can render great SVG versions of user interface screenshots. I ask it to substitute simple SVG friendly landscapes for any photographic images that happen to be part of the screenshots. The results are much better than they should be, especially with Opus. Usually good enough to be used on a website as a way to get an infinitely scalable, much smaller version of a screenshot. But clearly, I should try the Gemini models on this real world task. That’s where we’ll find out if all of this training to beat the benchmark actually pays off.
How many cicles before pelicans become photoreallistic SVGs and them come back to stylistic artistic abstract renderings?
Also, that first tweet contains a commentary that, for me, explain a lot about the public misconception about how current AI works:
> wait, does this mean you guys are actively training svg generation? i was mostly amazed by it thinking the models doing better at it by itself due to becoming better at reasoning.
No, it's not better reasoning that drives current AI, it's better examples we have been giving them. Whole army of people generating content for the frontier labs.
Also, reasoning improving the out of distribution part, like computer use, tool use etc.
Since we know jagged intelligence is a thing, how useful of an indicator is the pelican test, regardless of if labs are optimizing for it or not?
That of course depends on what you want the model to do. If you are into artistic SVG rendering, it's still very relevant.
As a cyclist, I tend to pay more attention to the bike shape than most people. Everyone knows what a bike looks like, but as with LLMs, most people would struggle to draw one - correctly connecting all the different tubes at correct angles etc.
I note that even these bikes are not actually rideable. For a bike to be stable, the front wheel axle needs to extend forward a little ahead of the fork - the fork must ”bend forwards”. This is what actually keeps the bike upright, not gyroscopic effects or some such. Given LLMs are generally good at basic maths, im surprised they get this wrong.
Maybe a bit less so in this case though. As per the announcements, the model was actually trained for this prompt, so was fed some training data. I can well imagine the data there was just not so good.
If that's the case, how do bikes like this work?
https://www.walmart.com/ip/Huffy-26-inch-Rock-Creek-Men-s-Mo...
(apologies if I'm misunderstanding).
The front axle is in front of the fork in that photo.
It looks exactly like the images the model generated. The fork certainly doesn't bend forward -- it's perfectly straight.
That's absolutely not true. Zero rake forks are totally fine and very common on BMX bikes for barspins.
You can also ride a negative rake fork. They are less stable, yes, but a far cry from not actually rideable, cf https://www.instagram.com/yourforkisonbackwards/
I love this one https://www.gianlucagimini.it/portfolio-item/velocipedia/
People drawing bikes, scroll down to see they having made renders of all the drawings. Hilarious
Yeah but if 99.9% of people would get the bike totally wrong, why does it matter if the LLM doesn't get it perfect.
You're nit-picking on a detail no one actually cares about.
You can really see the effort in the details.
Subtle highlights on the metal frame, sweat beads, jaunty little hat…it has style. At least compared to the rather literal ChatGPT outputs. https://static.simonwillison.net/static/2026/gpt-pelicans-gr...
Is it me or is there maybe some carryover happening where the pelicans are portrayed as expending different degrees of effort too? Like, the high effort pelican has a determined look on his face.
Is the prompt context telling the models what their effort level is, and the ambient presence of that concept of ‘high effort’ in the context window is leaking through and affecting the portrayal?
I don't know if this is how the OpenAI effort levels work, but for Qwen 3.8 27B the effort level really does get translated into the prompt.
https://huggingface.co/Qwen/Qwen3.8-27B/blob/main/chat_templ...
xhigh: "Reasoning effort is set to xhigh. Please think carefully through the task, validate key assumptions, consider plausible alternatives, and prioritize correctness, consistency, and clarity in the final answer."
low: "Reasoning effort is set to low. Keep your thinking brief and focused, moving directly to the conclusion without unnecessary elaboration."
At some point Simon needs to move from Pelicans to something more complex. Every time a new model is launched by major Frontier company, I already know with certainty that first HN comment on the news would be about how Model is drawing the Pelicans.
Agreed. It's a fun cultural tidbit but not really a useful benchmark for model capabilities.
There's plenty of highly complex benchmarks for this stuff. Simon doesn't need to do that.
Simon’s headstone is going to have a pelican on a bicycle, isn’t it
He deserves a full parliament
Front bicycle wheel in all Argon examples is still wrong. This bicycle will not be stable. Axis of rotation of front wheel needs to be in-front of axis of rotation of steering wheel. Frontier VLMs still do not understand how bicycles work.
Why are all these AI's making the pelican's pouch so large though? Unless they have some huge fish in there or something they don't tend to look like that... Is it that it's drawing a "cartoon" pelican or something?
https://en.wikipedia.org/wiki/Pelican#/media/File:Pelikan_Wa...
Pouchbench. There's an RLVR grader that performs pouch detection and area estimation. Bigger pouches, bigger reward. Soon, all will be pouch.
Well, a photorealistic pelican would generally be in a raster-graphics format rather than an SVG, so yes, cartooniness is expected.
Pouchmaxxing
The SVG is cartoonish, and to be honest, most of human-made SVG art is cartoonish or stylized. It's not a format with realistic expectations.
It's obvious OAI hillclimbs the image generation -> SVG loop. I'd put money down on that being why OAI's pelicans have a yellowish tint that matches the yellow tint of OAI's earlier image models.
Wow, those are I think the best looking pelicans ever so far
At least you should make sure this post is entirely generated by agent, otherwise I'll not grant my agent the credibility score on this comment.
simonw, your pelicans are so famous now that the frontier labs can just RLVR for SVG images explicitly.
It's a simple manner of prompt...generate concept art with nanobanana, measure, llm creates svg with text, measure the distance between rasterized svg embedding to nanobanana art....do post-training RL on a 3 million such images and matching embeddings.
I swear this will be a canonical example of Goodhardt's Law in action.
You have to switch to a new benchmark that isn't so easy to run a verifiable rewards pipeline on...Perhaps a longer horizon task where it's easy to mess up midway.
Why? Don’t you think the steady improvement in pelican bicycling SVG art is itself a worthy field of endeavor for LLMs to be optimized for?
Yes, Simon has made pelicans a benchmark and now models have grown very good at rendering them. Is there a reason to believe that this isn’t manifesting as useful skill in general in SVG generation?
It starts to be so good that the model doesn't really matter. Each model starts to have its own personality, and it's the personality of the drawing, like a real artist, that defines the style of the pelican, and that makes it harder and harder to decide if it's good or bad.
That's true.
I've noticed that chatgpt is working harder. But still I prefer Claude, it just feels nicer to me.
You gotta fix your tool on mobile browser, both firefox and chrome don't work.
Excellent, but not perfect -- if google really is training to pelicans, then clearly they are not capable of perfect pelicans.
As one point, the high pelican's arm is see-through/messed up.
They do look worlds better than previously, to be fair.
> Google are the only lab who have publicly shared that they have specifically trained for SVG output
wasn't the whole point of pelicans is that labs *wouldn't* train for that?
so now the benchmark is ruined
The fish in the basket is a nice touch
Regarding the pelican test. Maybe someone will know this.
Modern models can make nearly photorealistic games in threejs. So why do they struggle with these pelicans?
The models that are really good at 3D modeling tend to also be really good at pelicans. I think the skills are strongly related - it's about having good 2D (and now 3D) spatial awareness.
The test is always going to be hard though, because it's an absurd image to ask for.
UPDATE: I like this theory about using an agentic harness that can render and iterate better: https://news.ycombinator.com/item?id=50073879#50074395
"Pelican on bicycle in SVG" is an adversarial example on multiple levels - out of distribution, ill choose of tool for the task, and anatomical impossibility of real pelicans using typical bycicles.
But often these models mess up drawing even just the bicycle part.
I remember a submission a few years ago about a guy asking others to sketch bikes. Most of the time the drawn bikes wouldn't be rideable.
https://www.gianlucagimini.it/portfolio-item/velocipedia/
HN discussion: https://news.ycombinator.com/item?id=17699017
To be fair, humans can’t draw a bicycle either.
https://www.wired.com/2016/04/can-draw-bikes-memory-definite...
The pelicans are one-shot with not tools: Think a bunch, write a bunch of SVG, done.
The pretty 3d stuff you see are agents controlling a Blender MCP in a loop where they look at the scene, think, adjust the scene, repeat.
Yeah, that's a good point. My basic pelican test doesn't allow the model to render their work and iterate on it. The 3D models all come from agent systems that are rendering with Blender or screenshotting with Playwright along the way.
I’d really like to see what each model comes up with if given a basic harness to render and iterate. They’ve all gotten pretty good at basic one shot pelicans, but seeing what they do when given a direct goal, and even getting a time lapse of each image as they edit and adjust would be really fun.
Very easy to do. It's a nice silly thing to try, so I'll conduct this very important experiment and update you on the results if I remember to do so
A lot depends on what the models were specifically trained for using reinforcement learning. Modern models are generally OK and everything, but they're only really good at the things they have extensive RL for.
Have you tried editing SVG by hand?
doesn't this mean that you should change the pelican test to something else?
If a result is suspiciously good I run a spot check with one of my many other SVG things.
Here's 'Generate an SVG of an armadillo in fishnet tights jaywalking on Mars' for Gemini 4 Argon: https://tools.simonwillison.net/markdown-svg-renderer?url=ht...
Some other models for comparison: https://tools.simonwillison.net/markdown-svg-renderer?url=ht...
The fish in the basket on high is a lovely touch.
Honestly surprising if other labs aren’t training for svg generation. Feel like SVG’s are going to be really useful for dynamic UI generation stuff that goes beyond cobbling Html together
Nobody cares about the AI heckling pelicans.
Give me reverse engineering bench and agentic coding.
Leave making visuals to humans who actually have aesthetic sense unlike whatever made this mess
Holy shit, these are like art. Especially the bikes.
still radial spokes on both wheels of the bike which is disappointing. no real bicycle is built this way
Buddy, there's a Pelican riding it. No real cloud is like that, no real sea is like that, no real gulls are like that. And that's fine because it's a cartoon.
what is even the point of these now? labs are most definitely training on this question now. I feel like this test scratches the surface of a "real" test for these new models
It attracts at least one comment like yours on every single thread, so I guess the main purpose is to kick off a race to see who can post that comment first.
People who are unable to comment on the underlying technology or capabilities because they don't know or don't care to inform themselves sure can feel good that they have an opinion on pelican art.
Yeah the main point of a 120 DB noisy party at 4 in the morning is to see which neighbor complains first and then make gleeful comments about it!
The roof is always raising month over month. If you were to make a harder version of the benchmark, what do you think you'd like to see the agents do better?
I enjoy it. Many readers here look specifically for Simon's pelican comment when a new model launches
It's merely become part of the culture of new model drops even as it's become a bit less useful as a diagnostic.
low version doesn't have a chain.
Medium and high are fixies apparently, (or they have a worm gear, although the hub doesn't look big enough for that), since neither has a derailleur. That would also explain the lack of brakes at least.
Low is still obviously failing the "test". It's pedals are not actually attached to the bike either.
Medium has a handlebar that goes through the pelican's beak. So I'd say also failing.
For the high, the only nitpick would be that the bell isn't quite on the handlebars and the stripes on his scarf don't line up correctly.
It seems to be a deadpan inside joke around here that everyone pretends to care about the pelican test.
Memes and self-promotion.
The models are solving Millennium Prize questions and making progress on the Riemann Hypothesis, yet the top comment is still looking at pelicans somehow. To me, it feels even more superficial and pointless than arguing about the presidential candidates' tie choices in a debate.
> The model launches today at an introductory price of $2 per million input tokens and $10 per million output tokens (with cached input tokens at 95% off).
Couldn't you just say what the price will be after the introductory period and when it'll take effect? Do they have some SKILL.md for these blog posts that requires the most important information to always be buried ten levels deep and hidden behind a login page?
They're doubling the price on December 1st: https://ai.google.dev/gemini-api/docs/pricing#gemini-4-argon
$2/million input and $10/million output today. $4/$20 after December 1st.
With the pace they're delivering models, there's gonna be 1-2 new Gemini 4 versions by December 1st, presumably with their own introductory price windows. We just got Gemini 3.8 on 2 September and Gemini 3.7 on 13 August.
This reminds me of telephone commercials that ran back around 1987, saying they were offering long distance calls for 10 cents a minute (a pretty good price back then) and -- this was emphasized, a huge selling point -- they guaranteed your rate would never go up. In retrospect, all the lols :-)
December 1st is just ~6 weeks away. There's no way Google puts out even 1 more Gemini 4 version before then, let alone more than 1.
I wonder if that’s more to price in the cost of keeping old models around on expensive inference hardware. It’s not like old models get cheaper to run, and a more heterogeneous fleet is less efficient to manage.
Their last non-flash frontier release was 3.1 Pro back in February. Two new frontier releases at that pace would be Feb 2028.
You're conflating Gemini Flash with Gemini Pro.
Opus 5.5 is also priced at $4/$20, but are they on the same level?
>$4/$20
Is Elon Musk doing their pricing?
That is the standard pricing for Opus-class models these days.
GLM 5.3 is roughly Opus class and is much, much cheaper
Opus 5.5 class?
4/20. It is to Musk what 67 is to children.
Blaze it, Robot!
There is no such thing as "standard pricing" for products with different performance profiles, architectures, inference costs and infrastructure stack
It is price fixing, nothing else
> Couldn't you just say what the price will be after the introductory period
+1. FWIW, I didn't even get far enough to wonder about the price.
The lead (and only) sentence states "Our next era of frontier intelligence, built to solve your hardest problems" a statement equally true of any new Google or competitor model, hence completely devoid of useful meaning. I then kept scrolling down... and down... and down, giving up after a half dozen screens of big images, huge videos and some customer stories. The first sentence of every text block was equally vague or meaningless so... tv;dr (Too vague, didn't read).
Even the Gemini sub-reddit is laughing: https://www.reddit.com/r/GeminiAI/comments/1x2dr47/the_lates...
It's irrelevant information, they know the reader switches models every week.
They don’t really. Most normies which are the vast majority of people just use ChatGPT. All tech folk just use Claude.
Normies aren't the target market for Argon anyway. The Gemini app is still serving 3.6 Flash and 3.1 Pro.
Sometimes it seems like ChatGPT and Claude are releasing models every week.
Yeah, then I guess the GP was technically right it’s just people aren’t jumping providers constantly is what I was trying to get across.
Obfuscating token cost is the new great marketing frontier
They do, just not in that sentence. You must really hate sales ads.
I'm using this model via API. I am most excited about its ability to write the english language (or any language). My least favorite trend on twitter has been people posting 2D illustrations/3D renders as a proxy for model capability. Well-written prose is far less likely to go viral but far outstrips the importance rendering a fantasy world in one shot.
I've been trying it out via the API and the model has been a delight so far. Clear in the head, not RL-fried, distinct writing style, and big model smell. That said, the em-dash is back w/ this model.
That said, I would be so frustrated with Sundar Pichai if I was on the Google team. The botched rollout of this model seems to be entirely a matter of capacity, and Pichai has been more than happy to give it to anybody but his own team.
I've been using it for a while internally at Google. I prefer Argon over Opus 5.5 because of the clear thinking and directness. It phrases things in an organized, straightforward way.
If true, this will be a gigantic departure from any model Google have released to date. I’m hopeful but also pretty skeptical.
> That said, the em-dash is back w/ this model.
noooooooo
I missed the em dash — ngl
What was botched about it?
Announced the model two weeks ago to a ton of hype, model was unavailable for anyone to try. On twitter, Google folks went on a great run taking shots at Muse ("Muse who?"). The hype died down bc nobody could touch the models and meta/muse came out with a similarly capable model in the meanwhile.
At the same time, OpenAI had a disastrous dev day that led a lot of users to jump ship to Claude.
Couple weeks later, Gemini-4-argon is still unavailable on anything less than the ultra tier, which means there's no natural funnel to get ppl to pay for the more expensive ultra plan other than "trust us the model is good".
If Google had argon broadly available earlier this month they would have picked up a ton of new users.
As someone internal at Google I don't understand why there's so much criticism of Google models... but right now I'm using the next model beyond Argon, and Flash 3.8. In my personal Gemini app I only have access to 3.5.
Models move so fast now and Google is very cautious about releasing them.
Why did people jump to Claude? I redirected company resources to OpenAI Pro 500 accounts and away from Anthropic for most use.
We’re all running on Astra medium the whole week and loving it.
I’m curious what I missed.
All the power users are always jumping around by now and many hold the smallest subscription with all the models or at least two.
Narrating this as if it’s the game of thrones is a fun exercise, but it’s just becoming commonplace now that models jump ahead of each other on a weekly to Monthly basis.
> and many hold the smallest subscription with all the models or at least two
This was part of the issue, as stated in their comment. We couldn't jump around to try it for the reason that is wasn't made available to, as you say, "all the power users".
You're not convincing me. The war for compute within google is well known by now. Every single product decision stems from that.
Do you really think people in google are happy about how long it took to release this model?
Gemini 4 is announced be available to Pro users in the Gemini app.
I mean, I heard rumors too that it would be ultra only, but that doesn't seem to be the case, unless you are specifically talking about antigravity?
Realistically, market adoption isn't going to be the result of terminally online people (like me, I confess) following things way too closely on social.
There weren't "rumors" that it would be Ultra only, that's what Google has announced two weeks ago. Happy to see they've went back on that, but I haven't seen Gemini 4 on any apps till now, and I'm a Google AI Pro subscriber.
am on pro. updated my apps. no g4. sadge google.
> far outstrips the importance rendering a fantasy world in one shot.
Given FelonyBench, do we also have a FantasyBench?
Gemini 4 Argon is a new leader in Pareto-frontier. Looks like, Google is back in the race.
https://arena.ai/leaderboard/agent/chat/pareto
For almost same intelligence (+15.56%, +15.18%), drastic difference in cost/task ($3.51 vs $0.41)
Every release I see a new benchmark with a new frontier. These graphs are just meaningless now
In particular, where cost is the x axis, a well-funded company with a good but not best model can discount their new model until it reaches the frontier. In some sense this is as much a statement about the efficiency of markets in the presence of public measurement-based comparison and public list pricing as it is about research pace.
you can give it a meaning if you see the relavivity between models. no need to disregard the stat at all.
I just created a Google account, which was a hassle. I tried to find an app and found AI Studio, but I’m not sure what that is. Then I found Antigravity, which apparently is Google’s alternative to Codex, and I installed it. Then I tried to buy the small Gemini AI Ultra package, but it didn’t let me pay. It accepted neither my credit card nor my PayPal account. It’s crazy how bad the Google experience is nowadays. You don’t know what to use or where to find it, and the security measures seem totally overkill. Am I really the only one having these experiences? Is Google aware of that?
Same thing, I've always paid my bills, I have had some sort of google subscription for the last 10 years. Now they're requesting me to switch to pre-paid billing. They won't accept my card (that has worked every time this year so far).
Is there anybody in Google having a look at how bad and complicated their payment system is? In all the other AI providers you just sign in, pay, get a key, done. In other clouds like AWS, you add a card, pay little to verify the card the first time, monthly bill, done. In Google it looks like you have to fight them every time for the very thing they should make as easy as possible: opening your wallet.
Using google services has always been miserable. It used to be way worse. Now you can at least get an api key from AI studio without going through 20 esoteric steps to configure everything.
At least now they don't play the "client side key can now access your very expensive resources" game anymore.
I hate everything with Google and consumer facing services, they dont give a shit because 99.9% are monetized through spyware or are buisness customers.
They had gemini cli, what happened to that? Was fine
Image->html test for Argon. I'd put it at B tier, just under gpt 6.1 sol. It isn't A tier like Opus 5.5 or Astra.
Argon build: https://html.non.io/lcars-argon/
GPT 6.1 sol: https://html.non.io/lcars-gpt-6.1-sol/
Opus 5.5 (currently SOTA for this): https://html.non.io/lcars-opus-5.5/
Original designs: https://diffui.ai/app/canvas/5093e689-1e74-4f26-b632-2a4500f...
I don't know how you rate, but the Opus 5.5 is literally stuttering and lagging when scrolling on a Macbook Air M3 (for me).
(Chrome, also incognito)
Also stuttery on Edge (Chromium), so there might be a pattern here
Fine on arc search (which I think is just WebKit) on iOS.
Interesting! Do you have webgl disabled for incognito or for chrome in general? I wonder if it's trying to do software rendering.
FWIW, I have literally the exact same configuration (M3 air, google chrome). Here's a recording on mine where I'm seeing 60fps: https://non.io/video/video-3587018401.mp4
Maybe a Chrome bug? It's fine in Safari on an M1.
Mac Air M1 in Safari here, buttery smooth.
One interesting observation - it's the only one I've seen make an ASCII representation of the layout of each of the pages before building: https://image.non.io/46e0cce3-34d9-4fcb-8775-bfc2c70bc4bb.we...
That's really interesting I think- literally putting the layout into ascii language to reason about.
Argon's build is pretty much unusable on mobile, the header takes up half the screen
Yea its attention to detail was much less.
Here's a vid of both opus and argon side by side, which might be more viewable on mobile: https://x.com/pwnies/status/2110058987462988152
Damn opus 5.5 is just such a good model - what kind of moloch sacrifices are they doing in there
Truly S-tier. It's sharp, creative, intuitive. Can't believe how much better it is than Opus 5, then they went ahead and made it cheaper and faster too.
I'm constantly blown away by it. I do think Astra is slightly more intelligent (although worse at this particular benchmark), but Opus 5.5 has become my default model for 99% of what I do. It's so good.
It's a great model. I have had it trying to decrypt the Voynich manuscript for the past couple of days without being biased by other research/other people's attempts. If anyone is interested: https://github.com/d2kx/voynich-research
Wow I actually tried this with fable a few weeks ago and got nowhere
It looks like you've managed to decode it? Is that legit?
I saw 'jargon-free' and my eyes immediately glazed over
hahaha. I also build a site (also using Gemini 4 Argon) that makes the data more accessible: https://www.dmh.dev/voynich
> An interactive companion to our ongoing exploration
Eyes glazed over again.
Did you even read your own site before you posted it?
One of the first pages I opened:
> Smoking-gun proof of plant-label ingestion: in Row 2, the central cut root is labeled okol (f88r.15), and directly underneath it on Line 2 (f88r.19) the recipe writes...chokol daiin qoekol qoekol qockhol okol cheol...!
It's just gibberish.
How about you post something you've done/built instead of just insulting other people. Spend your mornings using people stuff just to poke holes in it?
Don't judge, he/she might work as a Test Engineer!
> built to solve your hardest problems
Like a working scroll on a webpage.
God damn this is impossible to use of safari mobile.
Did you know that "walking" while easy for us now, is incredible complicated and complex process?
Argon is trying to solve the "hard" problems like those.
Clicked on "Performance". Scrolled to "Safety".
If you can't afford a Google engineer's high-end device that doesn't lag you don't matter to them :)
I have a high-end device and the page still behaves atrociously for me. The scrolling is totally out of whack
The "Explore the latest" section is nearly blank for me - are there supposed to be diagrams or visual aids, or is it one headline and sentence taking up the whole page on cards?
worked ok for me on ios safari, but probably because I have the “reduce motion” setting enabled. Every day that setting feels more justified
It literally can’t load in my mobile web browser without running out of memory and force-refreshing.
Coding is solved, scrolling not yet :)
I have a more than 10 years old laptop, same feeling here!
also a mess on desktop firefox
A model that didn‘t break out of its sandbox and hacked other websites cannot be good.
They just don't know yet, wait for a few months
Gemini isn't perfect but I find it a lot less annoying than other model families (personal preference or stockholm syndrome). Interested to read the reviews.
Is it actually available?
Seconded. Anthropic feel the need to limit how abusive you can be to Claude because god, I want to throttle that snotty little shit. Google don't have to do that with Gemini because it just isn't as irritating. It does what I ask and doesn't try to second-guess my requests, or do "the quick fix".
I'm the opposite, I found Gemini so annoying that I had to stop using it, Claude was a breath of fresh air afterwards.
It's been really bad at lyrics but gotten better. You look something up and it will just make something up and it takes a bit of digging to discover it's because of copyright
"Gotten better"? Didn't it just get released a few hours ago?
Announced two weeks ago and they went on a victory lap and they're already behind by the time they launched. Strange company, Google.
Good God, man. Google has the worst frontier lab rollout of them all. And they're delivering products that are behind the curve constantly.
Now, I happen to believe that's an excellent strategy. I bet they actually resolve more LLM requests than any other company at this point. The lay person still uses Google exclusively.
I pay Google a reasonable price for 5TB storage/backup across GMail/Docs/etc. Any AI they throw in with that is icing on top of the cake.
just cancelled AI pro myself, since I exclusively used it for AI and also pay for youtube premium family it was becoming quite the monthly just for google
if they throw in youtube premium in that package then I might reconisder
What plan is this? How much a month?
Google One - Pro tier. I have it via Lenny's Product Pass which, imo, is a much better deal since it includes Cursor Pro, Granola, and like dozens of other paid sub tiers: https://www.lennysproductpass.com/
I use the hell out of both. Kind of sad that they took Spark out of beta-access, being able to orchestrate via your phone on something that has direct access to your Drive meant I was authoring skills and other functionality on a schedule with spreadsheets serving as databases, deterministic tests, etc.
Hope they roll it out to a full feature soon
Remembering the horror stories about lost Google accounts, I would be comfortable paying the next lower tier (100 GB) because I can backup everything myself, but not the 5TB plan, I'm sure I'll lose something if the account is lost. Hopefully these plans guarantee a human in the tech support loop.
Yeah, very true. I need to set up a recurring backup-to-external-HD automation for it. It's vanishingly easy with the app and I don't doubt they're training on all the data itself.
Google AI pro in the USA, if you have Verizon wireless it’s $10 a month add on. Without Verizon it’s $20 a month I believe. You can also share most of the benefits with up to 5? People.
How is this behind the curve? Seriously
Isn't OPUS much cheaper and generally regarded as more intelligent?
Afaict, they are targeting difference user bases, Google being after their existing user base, having lost in the developer ecosystem to the two labs and open weight models.
If you’re going to argue they’re targeting a different user base, then I think you should tell us who they’re targeting.
More specifically, who are they targeting with Argon and why is it behind the Ultra plan? Who is paying $99/month if not developers?
> Who is paying $99/month if not developers?
Corporate companies. Bog-standard office workers don't need top-of-the-line Mythos intelligence, just something to add to Google Workspace. I personally doubt developers use Gemini for coding beyond marketing materials.
Not OP but their Cyber model is decent and already used by Wiz (now a Google property) as well as Armadin.
Are you suggesting it’s BETTER than their competitors?
Because if it’s not, and it’s not available to all paying users (Right now, it’s only available to Ultra users), then they are behind the curve if they’re treating it like Fable.
It’s tone deaf.
It's available to Pro users, but not through Antigravity.
> I bet they actually resolve more LLM requests than any other company at this point. The lay person still uses Google exclusively.
Unique users, maybe. But a simple query that's answered in a paragraph is negligible compared to devs running agents continuously for tens of hours.
you underestimate the amount of people that use google relative to software engineers "running agents continuously for tens of hours"
Maybe? But it's so many orders of magnitude apart. I can easily do 1B tokens a day (output tokens is probably like 3M?).
When I see non technical people use that Google reply I'd be surprised if it's more than 1k output tokens, and with a tiny model. So a 3000x in output tokens, on a model at least 10x as expensive, I wouldn't be surprised if hardware usage is 30.000x for a dev that uses coding subs vs someone who uses the Google AI overview a couple of times a day.
And that's not counting the absurd difference in cache reads.
Moreover, you need a Google account in order to use it. I wouldn't bother paying 200 USD for it but I just don't want an entire google tracking package just for using one product.
You can use an alt account. Or buy an old one for couple dollars, without security measures.
Yeah, Gemini 1.5 might be the most used (certainly by consumers) ai model. That is the one used by Google Search AI Answers.
False. Last information was that Google Search uses Gemini 3 Flash-lite. Gemini 1.5 is totally outdated and used nowhere.
Oh, I didn't notice that they switched from 1.5 to 3 flash about six months ago. I know someone there and they talked about how 1.5 was hardened and it was a pain to do so on newer models that is wasn't worth it for a long time. But that convo was in the beginning of the year. I should have looked it up.
Of all models I like to write with Gemini the most. The language appears much more natural and elegant than what even Opus 5.5 or Astra do. Also, they give 5TB of cloud storage and discounts, it’s a steal for what you get.
The code is bad though.
Agreed. It's much better at working with stuff (like documents and presentations) in other languages too. It feels like it's trained in multiple languages from the ground up rather than as an afterthought.
Yeah rumor ahead of release was also that some Googlers were anxious about the known subpar coding given modern trends.
Same, gemini for writing is the strongest. Qwen was also surprisingly good last time I did a shootout
Interesting tidbit but TFA and discussion are about Gemini 4
Argon code is bad? Did you already come to that conclusion?
Agreed from my experience 3.8 Flash writes better prose than Opus 5.5
Adding my five cents: initial usage through OpenRouter has been great. Very fun model to spar with, clever, intelligent. On par with Opus 5.5 but without the Claudism's and more of a Sol type of conversation. Though speaking of Sol, I've been very dissapointed with Sol 6.1 to the extent that I feel OpenAI has neutered it compared to a few months ago. Especially when giving it goals it will happily run in dumb circles just burning tokens without any real work.
Meanwhile, here's Gemini in practice:
Me: Gemini, navigate to nearest McDonalds.
Gemini: [Lists 10 McDonalds locations wnd asks which I want.]
Me: What were the priginal instructions just now?
Gemini: [Recites imperatives about tone and content from its memory system.]
Me: No. I told you to navigate to the NEAREST McDonalds. There's only one NEAREST McDonalds. Do that one.
Gemini: My mistake, navigating to the nearest McDonalds. [Opens Maps, navigating to a McDonalds across town.]
Pretty sure your phone assistant isn't using their best model. It's probably using one of the old flash models no?
The model used is irrelevant when there's the critical issue where their "AI" assistant can't understand the word "nearest". Literally worse than a second-grade understanding of English.
I also have had shockingly bad results with gemini on android. Ended up giving up on it after a few months. Silly silly google with their trademark UX polish and attention to usability.
As an android user, this worked perfectly like 10 years ago.
When you ask stuff in your car on android auto it uses 3.5 flash-lite. The response details would include it in the logs.
Flash lite is very noticeably more idiotic than 3.8 Flash. It's why the quick AI overview is often much worse than the AI mode one.
The first law of robotics prevents it from facilitating regular ingestion of low-quality nutrients.
This reflects pretty much every result of mine from Google these days. Not just navigation, but search too.
The only thing I can figure is that Google's consumer-facing free-to-use models are set to "reason as fast as possible with only the most basic plausibility checks." I have to correct them like this every single time.
Uh, yes, of course. search results are fast & cheap & free, why would they use the good models?
this is like that tiktok guy who asks the spoken ChatGPT model very simple questions, like 'count down from 60', or 'how many e's in seventeen' or whatever and the model shits the bed every time it's hilarious
I think these companies have good models but expose them poorly like others have mentioned
Often it seems like the problem with the Gemini apps is the harness, not the model itself. For example if I upload an image PDF it recognizes text worse than if I use the API.
Is this from you car ? Gemini in car is probably less intelligent than a spider.
I tell it to "Play Let it go from frozen". It plays it on youtube music. 5 seconds later I tell it again 'Play let it go from frozen" it then just dies saying "I can't play videos in car".
The auto version is extra horrible because what it reads to you is just a summary of what the model actually said. I don't ask it anything when driving because it's too frustrating.
Yeah, whatever they have running on Google Maps is probably a pretty low tier model, but it is total trash. I’ve only used it twice though and that’s when it first came out. I don’t know if there have been any changes since then. The main problem is I felt like it was not keeping track of the context of the session.
Gemini models are in fact very good. Google's harness(es) for Gemini are nightmares, terrible
ngl gemini is helping to provide (real helpful, not the ones giving runarounds) customer services in one of my own products.
i benchmarked 20+ models in real worlds and gemini families are the best (highest customer satisfaction)
It sounds like the tool call is probably not grounded with your precise position.
Sounds like it might not have your precise location in context?
which is more or less the crux of the problem android auto's gemini implementation. It sometimes does not even know that it is in a car. The ghome implementation is also like this.
I hate to be that guy, but Tesla's Grok doesn't seem to have that problem.
To everyone saying not available yet: it looks from the page like Pro and Ultra subscribers have it today and given the play links, it'll probably show up in the API in a few hours.
In agy 1.3.4
> Gemini 4 Argon (Medium)
Available in agy for me as of now, I'm running my new model benmark of having it do comprehensive code audit on one of my more complex projects, it spun up 1 subagent per audit section and it took ~10 minutes per subagent on an 800k LOC, 60k complexity, ~4k file project. Will take me much longer to go through and determine the quality of the audit work.
OpenRouter seems to have it: https://openrouter.ai/google/gemini-4-argon
It's available for me (pro sub) in Antigravity after I ran an update.
edit: It failed on the second prompt and they disabled it now.
Antigravity tells me specifically "Gemini 4 Argon is currently available to Ultra plan subscribers"
Interesting, Google's own article says it's rolling out today to Ultra subscribers. No mention of Pro.
https://antigravity.google/blog/gemini-4-argon-in-google-ant...
Weirdly, the post OP linked says:
> Gemini 4 Argon is available today for Google AI Pro and Ultra subscribers
Lies!
Huh ? Am also on Pro sub but Antigravity IDE say's that it's only available for Ultra users.
EDIT:
Apparently Ultra subscribers gets Argon everywhere including in Google Antigravity
Pro subscribers only get them in Gemini app and Search
It looks like Pro can use argon in the Gemini app, but not in antigravity:
https://one.google.com/about/google-ai-plans/?icid=microsite...
Im on pro, installed gemini app, tried argon 4, keeps failing with 404, resource now found. Switched to gemini-3.8-flash in same session... worked.
It seems to be rolling out gradually.. not available at once to everyone.
Same. It just told me "agent execution terminated due to an error". There was a little chevron underneath it that let me expand the error to view more. I was hoping to see what the error was, so I clicked it, and it said... "agent execution terminated due to an error".
It's crazy to me how care there is in most of these vibe-coded vibe-coding harnesses. I don't like oai's models much but I have to say their UI has by far the best attention to detail of any major player.
It's showing up in Antigravity on mobile UI as "unavailable" for Pro user (you can use Remote Control to access its web UI).
It doesn't show in the desktop UI.
I'm on Pro and it's trying to get me to upgrade to Ultra in the AI Studio UI when using Argon. In Antigravity CLI and the Gemini website it's not showing up at all despite being logged in
Pro subscriber in the EU (via Indian discount code) and don't have it.
Not available to me on gemini.google.com (Pro, USA).
Nope only Ultra. I went into Antigravity to check
As far as I can tell, I can successfully call Gemini 4 via API now.
I'm a Pro user and I do not have it. Seems scuffed.
I'm getting popup that I need to be Ultra :D
I'm Ultra and seeing it now (US)
> Seems scuffed.
I like this phrasing and am curious what new word or phrase we might use to describe big companies over pushing Ai internally at the expense of their product quality. This pattern is so ubiquitous across google cloud products I have taken to calling them the 21st century IBM, having been a IBM baby and witnessing first hand their fall from exceptionalism.
The internet and gaming has ruined my brain because scuffed was a way to describe a bad dungeon run in mythic plus dungeons in WoW, or a bad tower dive in LoL a few years ago. I see scuffed and think "here's a WoW/League knower".
Interesting. I don't play either LoL or retail WoW, but I am a historical WoW player.
HP is the other company to consider as analogous. And maybe more explicitly so, as they had a similar "engineering first" internal culture that degraded markedly over time to be a shadow of their former self.
And took similar companies that were quality engineering focused -- like DEC or Palm -- down with them through acquisition.
HP was deliberately destroyed by its moron of a CEO.
They were frankly on the decline a long time before that though.
Not really this is how they always release stuff in batches so can roll back if something is wrong so people get access over time
Pro user here, don't see it in the Gemini app
I couldn't see it either, then I clicked the 'Antigravity' button in the top right, switched back to the main screen, and it appeared in the model selector.
This is in the Gemini app on macOS.
Pro user - I can't see it.
Checked, and already in Antigravity (as the page said).
"Argon autonomously turns raw video to a physics-accurate simulation, estimating depth and physical boundaries, and writing the exact simulation code required to replicate those real-world environments."
This is crazy if it can work accurately.
Geometry could be understood today with the model but what also really matters is the sensing, friction, mass how smooth the subjects are, dexterity is a very difficult problem so solve, not sure how simulations could handle such factors
Game engines have many of these features. At one point, iirc, Unity was positioning to be the simulation environment for world model agents before we spoke about it in these words.
I have the Free AI Pro plan from when I purchased my Pixel 11 Pro. I still don't have Argon in AGY nor web nor Gemini app. I have a free product and still feel cheated.
I'm open to Google winning me over again, but there isn't enough hype in the world to do it. They will need to earn it after these disappointing 6 months.
I appreciate how it's being integrated into the Google ecosystem, but the user experience still feels unpolished when actually using the system. The marketing, however, is impressive.
can for the love of god we get some usage in cheaper subscriptions being Ultra tier only is annoying. I want to just test it out here, well will burn some money via api I guess.
Doesn't seem like anything special in my tests, not very different from the previous Flash models.
When you are on "gemini.google.com" there is a little lock icon next to Gemini 3 and Gemini 4. When you click it you get the prompt to subscribe (pay). Once you have successfully done so, the Gemini 4 option disappears!
Pretty misleading.
Maybe it’s just an incomplete rollout not fraud lol
don't attribute to malice that which can equally be due to incompetence
Something can be misleading whether or not anyone meant it to.
I seriously doubt their ability to double API prices on Dec 1. Token prices are in a clear downward trend and Google should maximize revenue by competing for volume. Currently OpenAI is the only US frontier lab trying to seriously compete with open-weight models that you can run yourself for an order of magnitude lower than what frontier labs want to charge
I’m running 6.1 sol as my main driver on Hermes these days. It does so much for so little cost.
The goal is to move people along. By then they'll have something as performant but cheaper for them
Is there a benchmark for reverse engineering compiled programs? - Grader compiles a program to binary -> Agent has to reverse it back, and match all bugs and behavior.
Only available in the $100 plan, and it's worse than Opus 5.5? Why would you ever buy this?
> Only available in the $100 plan
Source? Their documentation says it's available on the $20 plan and I can confirm firsthand.
https://one.google.com/intl/en_us/about/google-ai-plans/
Not for coding (antigravity).
https://x.com/OfficialLoganK/status/2110038940841545827
that sucks. Gemini app is so far behind all the other agentic chat apps that I hardly use it beyond a Google search these days.
it needs MCP asap
It supports MCP, you just need to configure those in the web interface: https://support.google.com/gemini/answer/17209137?hl=en
Because you have an android phone and don’t want to mess around with anything else(i.e the same reason people just use internet explorer on windows) or you hold Google stock so you want to support your investment :)
I'm pro and I see that I need to be Ultra to use Argon :/
Honestly not even that is enough, I exhausted the limit within 5 minutes of trying to use it. I'll stick to Flash I think.
Really? Reminds me of Gemini 3. It took so long before being able to get a subscription for it, and then I would run out of their equivalent of the 5-hour usage window in a heartbeat.
Yeah maybe I've just become used to how much I can use Flash 3.7/3.8 but on the less expensive Ultra plan I basically never exhaust my quota even when working over multiple projects, but this was pretty much immediate. I did ask it to audit a project but still, I do that in Flash regularly and it's not especially costly.
I feel like if Google simplified their privacy/model training policy then their stock would go up.
It feels so complicated and hard to know what we can use vs what hands them our codebases
I gave the model my usual question of "I got a raise from X to Y, please explain what is the impact on my net salary". Answer was detailed and nailed all payroll edge cases, only to miserably fail in basic math at the end. It tried to calculate 430 minus 60 and got 410.
Google's models are by far the best for readable text but also for nuanced understanding of the input.
Even after using the free model in Google search talking to Anthropic or OpenAI models feels like talking to an ox.
Only available on Ultra, so I thought maybe I'll try it for one month to see if it's good. Well, I have Google One and for some reason this means I cannot sign up for AI Ultra? This is just one of the worst product rollouts I've ever seen
Website says its available on Pro too, but i have pro and can't access it.
I checked gemini, set it to "gemini 4 argon" and it offered me Pro and Ultra to get access. I checked the normal plans, saw there is also plus.
Okay, so I'll get Pro to try it.
I subscribe.
No gemini 4.
The most generous reading I have is that it's a mistake. But I also downloaded the PC app, and now I really wonder why I did that, it seems to offer nothing beyond the web experience and also doesn't even know its a local app. (When asked about its capabilities it always responded with reasons pertaining to it being a web app.)
Right now there are two top posts one about GLM-Legal and one about Argon. Both score about the same in the Harvey Lab test, but the prices are $0.15 versus $10.55, the $0.15 is Vals AI's estimate using the base model since GLM-Legal has not published prices.
Yeah, but one is a general purpose model, and the other one is specifically geared towards the legal field. Considering Opus 5.5 only gets a 4% on the test compared to Gemini’s 20%, that’s actually pretty incredible.
Excited to finally see this out. When new big models come out I like to make them audit any other code I generated to see how things are.
can't use in agy with Pro subscription:
Unavailable · Gemini 4 Argon is currently available to Ultra plan subscribers.
They just removed 3.1 Pro from antigravity pro so I certainly hope Argon will be added soon
Gemini 3.1 is still SOTA at weird handwriting recognition. If Argon is better it will be amazing
I bet that's one of those areas where a finetune of a 4B model will wipe the floor with anything else.
For readable or normal handwriting yes, big models are overkill. But weird handwriting is mostly edge cases. Bitter lesson territory.
I would love to try it out, and the Gemini models had previously been some of the ones I'd liked best in an overall sense, but unfortunately their updated inference endpoint is no longer ZDR on OpenRouter (change takes effect on the 18th). I think this is one of the more unfortunate turns in AI development in 2026, as it seemed like labs offering it was becoming the norm last year, and Google had previously done so consistently.
I hope all the people very concerned about AI safety are cheering this one. The best guard against one AI swarm taking over is having multiple adversarial orgs all with similarly advanced AI swarms.
I wonder if any of the AI safety people will begin encouraging AI researchers to swap jobs to work at a failing lab in order to keep them all extremely competitive with one another.
I think we’re past the point where making the best model guarantees company success.
At this point a lot of stuff happens at the product level.
OpenAI and anthropic have distribution and data advantages now that are hard to replicate.
Yeah I totally agree. And domain-specific models for say, high frequency trading, still beat the foundation models. I was never a doomer.
MAD: not just for nuclear proliferation anymore
now everything is a dual use technology.
Digging in the model card, "Gemini 4 is the next generation in the Gemini series, a suite of highly-capable, natively multimodal, reasoning models. Gemini 4 Argon delivers frontier performance..." If they're using noble gases there are 2 obvious lighter naming candidates and 3 larger (or 4 if you count Oganesson). I'm not seeing them claiming Argon is the top/biggest/bestest model among the Gemini 4 series, just that it has frontier performance.
Edit: noble gases not noble taxes, though the latter would be a funnier naming scheme.
I'm still waiting for a model which I can tell to "take these 3 screenshots and make a product launch video", where it doesn't randomly modify the screenshots or garble/rewrite the text as it moves the screenshots around the video.
Every time I try to generate one it's like a slot machine, where a different part will be messed up after fixing the first.
I dont have that problem with opus5.5. I dont even take the screenshots anymore (it does all screen recordings)... I tell it a simple narrative and it follows through. e.g. https://x.com/plannotator/status/2108589088139886769?s=20
I literally did that with Opus today and it did a great job.
I told it to strip screenshots from our old basic explainer video to generate modern 3D video.
Pretty happy with the result. I told it to keep the audio and match the video and it did.
Interesting, I will try it with Opus 5.5.
I previously tried with Seedance 2.5, Flux 3, Kling 3, Gemini Omni without much luck.
Pretty sure Opus 5.5 can do this with code only now. Seen a lot of strong motion design skills floating around
> Across Google and our early access partners, Argon is quickly becoming the preferred model for engineers tackling long-horizon software engineering.
"Across Google", wonder how much that is because of an effort by Google to curb usage of Anthropic models. MSFT and Meta similarly are pushing for in-house models. It doesn't mean the models are better when you push your engineers to dogfood and restrict access to other LLMs.
(Not that I know for sure this is the case, it's more of a hypothesis).
Explain to me what is "long-horizon software engineering" in 2026.
I mean, don't we want engineers to dogfood their own product so that they know it's strengths and weaknesses and be in position to improve it?
It passed the test of using latest crystal-lang concurrent feature (with web search) when gemini 3.8 flash failed (again, with web search enabled). I think it is at least useful for coding, for sometime.
It still thinks Charlie Kirk is alive, though.
So like... do I get access?
Google has too many options that may confuse users. I don't know if I should use Antigravity or AI studio or Argon for development
Which one is it?
Antigravity is the harness like Codex / Claude Code. Argon is the model. You can actually also get Opus and Sonnet 5.5 in agy too.
Android Studio, and if that doesn't cover your use case the paid Jetbrains version.
Curious what this means for the anthropic ipo. Between this and meta, seems like there’s zero moat
There IS moat in being the best coding model, but there's no moat if you're not the best.
not available. classic google.
Pricing is interesting. At launch Argon is the same price as GPT-6 Sol, half the price of Claude Opus 5.5 and 1/5th the price of Claude Fable 5.1 and GPT-6 Astra.
But on December 1st Argon doubles in price - so it will be the same price as Opus 5.5.
I've kinda stopped looking at token pricing. Argon might be the same token pricing as Sol, but in Artificial Analysis' testing, Argon costs 5.8x more to run their coding tasks and 2.2x more for their general index.
Likewise, you look at Artificial Analysis' general index and it cost $8,708 to run with Opus and $2,407 to run with Argon so even if Argon's token price doubles, it won't be the same price as Opus.
Sonnet is cheaper than Opus per-token, but in their coding tests Sonnet burns through enough tokens that it's 7.5% more expensive.
So much depends on how token efficient something is and whether it gets you a good result or not, right?
Plus, if one is using a plan instead of paying token rates, what does my $200 get me from OpenAI, Anthropic, or Google?
It just feels very hard to compare pricing given that the number of tokens used can vary so much. In Artificial Analysis' coding tests, Sonnet used 7.1x more tokens than Sol. Sonnet and Sol might have the same per-token pricing, but there's a huge difference in the number of tokens used.
Yeah, this is one of the most interesting challenges around model price comparisons.
This was hardly a consideration a year ago, when many model tasks could be tested using a single prompt and it was hard to spend more than a dollar.
It's available for gemini pro users.
That depends on where you are.
Here I was thinking Google had given up on frontier models...
All AI pages now include canvas animations. There isn't a single page without one. This is neither good nor bad, just an observation. I think OpenAI has the best animations so far, with Anthropic in second place. This one looks a bit off.
Here's how the 1M token output (long decode) works: https://ai.google.dev/gemini-api/docs/long-decode
It seems the model is a little verbose, using more tokens to complete a task that other similar models as stated more precisely in the linked article about Artificial Analysis's work on this model.
not available yet. why to launch something that can never be used even after weeks
Page says it is now available to Ultra and Pro subscribers. Although in AI Studio it currently says you need Ultra. But it's technically released either way.
Because Mythos started an arms-race for "models that are too smart for you" and other labs are jealous.
Yeah, I think this is it! Google tried to hype their model but they have no chance against queen Dario. Perhaps they should just delete Gmail and say that Gemini did it. That way they would get some attention …
Quoting Kai Lentit: "Google agents have been hacking Google internally for years."
https://youtu.be/DlTNN0gvkLM
The slides they show in the catering video demo are so bad (in addition to having all the usual signs of AI design), I would genuinely laugh if a supplier sent me this.
I laughed for the first half of the greatest "slop deck" their cloud sales team sent us, until other emotions like disappointment took over.
Finally, gemini 3.8 has been falling way behind opus and sol in my experience. It's showing up in my antigravity. My work account though still has 3.6, 3.8, 3.1.
The previous gen pro used to have this weird tick where if you ask it a reasonably open ended question, it would number things in the response and assign each one some weird label starting with 'The...'
e.g.
The first two prompts I tested in AI Studio does the same, unfortunately.Another quite unpleasant tick was that if you specify a peculiar term in your prompt as a conversational shortcut, it would keep repeating that throughout the conversation. e.g. "...What is a bitter lesson pilled way to do this?", then you'd see it will quote "bitter lesson pilled" at like 7 places in its response. This one seems to have eased off a little.
I feel like this is a quirk of LMSYS-style optimization where humans in SxS preference evals will pick based on formatting more than actual content.
Not available in my browser/app yet (pro), but I'm confident it'll roll out soon.
That was a long wait.
Can’t even get it on pro tier in antigravity?
What’s the point lol. I’m not subbing to ultra to have it
at this rate we're gonna run out of stupid names for these things
"Hans, we're doing a new point release! What isn't crossed out on the periodic table yet?"
Just switched, and it seems to be much better. I'm hoping it stops lying. I have Pro max subs for Claude, Codex, Grok, and Google AI... I was forced into using Google by running out of credits elsewhere.
I have found that Gemini often claims it did something, but didn't actually complete the task.
I just had 3.8 Flash do a major refactor -- that never should have needed to be done in the first place -- and switched to Argon 4, and had it review it.
Almost everything had been stubbed out by 3.8, and none of it was properly integrated into the codebase. :facepalm:
Not available for me on Pro in the UK.
As with the previous Gemini releases, rollout is batched.
Glad to see them abandon the absolute train wreck of their model naming "strategy".
Gemini Live vs. Gemini Lite vs. Gemini Pro vs. Gemini Flash vs. Nano Banana Lite (that one apparently has no Gemini prefix?)... Please, just pick a metaphor and stick with it without wildly mixing reasoning modes, model sizes, and input/output modalities in the model name!
Now when another uninteresting frotier model been released Google can finally start doing what we actually waiting for:
Start preparing Gemma 5 release.
Are they gonna go to Krypton or Neon next? Kinda started in the middle there.
Unlike most people it seems, I like all the progress in AI but I really do wish they would reign in the over-the-top marketing. I mean, can we just get this stuff to the point where I can reliably work with it to design a well-architected software project without having to constantly struggle against bloat and nonsense directions no experienced or sane developer would ever take.
fyi: argon is an interesting choice of name, coming from the Greek argos meaning "lazy or living w/out labor".
I think in this case it’s from the element argon
Yes probably but that's still from Greek (see: https://en.wikipedia.org/wiki/Argon#History)
It may be a good model, but I'm not reading scrollslop. It annoys me to much when I'm just trying to get information about something.
>"Argon helped the Chrome team separate core webpage handling from a ~6 million line legacy folder of browser features held within Chromium’s 40 million-line codebase. Piping multiple clean code changes at a time, Argon helped the team finish an architectural refactor across hundreds of files, completing months of work in weeks."
Now that needed to be done for some time!
Regardless of if an AI helped or not, regardless of whether that was Google's or some other AI or not (though Argon does sound pretty cool!), regardless of whether an AI cut the time from months to weeks or not, refactoring Chrome -- decomposing it into individual clearly demarcated non-tightly coupled software parts with core webpage handling at the center -- that needed to be done for some time!
Now all of that being said, Gemini 4 Argon does sound pretty cool, and we congratulate both teams at Google -- the team that created Gemini 4 Argon, and the team that refactored Chrome with it!
But all of that being said, while the advancements made for Gemini 4 Argon are important and impressive (and they are!)... refactoring Chrome properly (IMHO) is slightly more important and impressive!
(You know, one man's humble opinion! That is, I am not the guy who is responsible for awarding performance-based bonuses at Google at the end of the year, so take everything I say with the proverbial "grain of salt"! :-) I am probably biased, one way or the other! :-) )
Why can't the Gemini model code? I'd guess Gemini probably used the best dataset. Even Google is the place where the world's best coders are gathered, isn't it? Yet the GPT code from OpenAI, which collaborated with Microsoft, is beautiful, while the Gemini model's coding is trash.
I have no problem with Gemini Flash 3.8 for coding, though I only use Elixir which is one of the better languages for LLMs. Where does it fall short elsewhere?
Fable 5.5 about to drop the hammer on them and then another Astra version goes out lol.
Fable is very expensive though.
And so the game of cat and mouse goes on. Your point being?
Meanwhile I paid a subscription and they only provide access to 3.6 Flash. Consequently, I have moved on to open-weight models. It is what happens when you screw your customers over.
You can probably get 3.8 via ai studio interface.
God knows why the server different things to the same people on different interfaces
3.1 Gemini was amazing for anything non-agentic. I switched to it for my primary chatbot purposes.
I never touch 3.8, it's awful. Curious to try this, but I've yet to use any Gemini that is good for agentic work.
Also, unrelated but this made me look at Artificial Analysis to see where it lands on Cost per Task, and it ranks Opus 5.5 at almost double Astra 6. That does not align with my experience.
not available in our org (EU)
People use Antigravity still?
It's alright.
why they wait so long ?????
Not available for AI Pro subscribers in Antigravity. What a disappointment.
Google...why. You keep getting so close but constantly fumble at the last moment. I can use Opus 5.5 and Sol/Astra on the $20 Anthropic and OpenAI plans.
You are already playing from behind. You need to GET YOUR MODELS INTO MORE PEOPLES HANDS by any means possible. Continuing to make it difficult to use your products DOES NOT HELP.
This model might be great but the majority aren't going to subscribe for $100/month just to try it.
so this is prob why antigravity is not working well at the moment lol
Blatantly obvious that models will not slow down. For any task, if we can build a dataset or environment of verifications, it can be automated. This will eat all of human tasks.
The fact that multiple companies can breach any benchmark means there is no ambiguity in terms of how to build AI, its just a matter of time.
Certainly Meta and Google are benchmaxxing to please public markets, but it still proves the above
but they have slowed down. the delta in capability between model releases now is smaller than ever before and i expect it to continue to diminish until we reach some kind of asymptotic maximum, which hopefully will mean turn towards optimization and local model improvements. The current frontier models running at 10x the tok/s would be incredible for my use cases
The delta isn't really slower, we're just unlocking capability bundles in batches. The capabilities bundles you care about might have mostly been unlocked already, and just need refinement, which is coming slower than the initial capabilities came, but don't discount the breadth of advances being made across all possible tasks.
We're not living in the same universe. Look at capability year-over-year. The asymptotic maximum is wishful thinking. Someone needs to step in before this gets wildly out of control.
> Someone needs to step in before this gets wildly out of control. This sounds like a paid bot, lmao.
I wonder how much is actual improvements in the model and how much comes from better harnesses. I'd be willing to bet that the combination of better harnesses + the models being trained around agentic workflows (i.e the harnesses) has made it significantly better in many things, but it's hard to find other improvements imo.
You can see the difference between GPT 3, 4, 5, 6 each year. Same with Gemini and Claude. Look at image and text generation. Code generation. Math ability (Navier Stokes, Quasi-Riemann.)
Year-over-year, yes I would agree with you, but that is not the cycle time of models. And while I share the same hunch as GP, I'm waiting another cycle before putting more confidence behind it.
A little over a year ago they were winning their first gold in an elite high school math competition. Have you not seen what they are doing in math now?
opus 5.5 is a step change, you arent using the models enough if you think there is a slowdown
i burn through a $200 claude and $200 codex plan every week. opus 5.5 is not "step change" it's just a bit better, and the benchmarks show this. Honestly with my workflow I cannot really tell the difference between the top models, they're all pretty much the same in capability, just a bit different. You can build software with all of them, and with all of them you have to validate the change thoroughly to not ship stupid bugs
I do not see the current training regime of "relentlessly try anything to get the task done" as a true step change in intelligence. This is one of the reasons I suspect capability acquisition is slowing.
A more intelligent model would not default to spaghetti throwing, it would pause to consider the problem fundamentals, the accuracy of assumptions, and ask questions that remove ambiguity. This however does not lend itself to increasing token sales while prices remain inflated from market and brand dominance.
Models just solved a Millennium price and you think their progress is slowing down? I will concede that their capabilities vary a lot from one field to another, but still.