There were so many memes about “Gemini Hydrogen” after they released Carbon. That they actually released a model named Hydrogen makes this very funny
But a new Flash model is very welcome. Argon and Carbon are very slow. In the time it takes for them to complete a single task, Opus already managed to complete multiple things. They’re so slow, I started using 3.8 Flash over them. (A lot of it is also because of how terrible agy is. Before doing any task, Gemini feels the need to hover up a bazillion files into its context when the task is something simple like “update the docs of this controller”)
I'm experimenting currently with reducing the hoovering aspect in my configuration of antigravity, on the other hand at least once when I thought the model was "too deep into sauce" it turned out Argon deducted a thing I considered completely unrelated was cause for crashes I asked it to debug.
Google’s segmentation here is keeping them from using their existing account dominance to eat away at OpenAI and Anthropic personal subscriptions, tbh. Adopt a lower usage limit for Carbon on the $20 plan and it’s suddenly insane value.
Given you get YouTube premium, file/photo storage, and $10 a month back in GCP credit which you can use on more tokens (including carbon). It already feels like great value to me, that said how much you get out of it will depend how deep you're in Google's ecosystem.
YouTube Premium is only personal so it's useless for me I'm already paying for family plan, and what do I need 20TB of cloud storage for to update to 100$ plan - basically it's a useless bundle for me not a value proposition beyond the 20$ with their usage limits I saw.
Same for me. I find myself using Google’s $20/month plan, plus running really good local models with occasional use of Deepseek Harness with their flash model, occasionally a $6/month Alibaba subscription for their strong qwen-3.8-max model, and about every third month getting a Anthropic subscription if I need more access to Opus.
This is a lot of purchasing complexity for an old retired guy like myself!
I wish Google had finer grained $30/month and $40/month subs: then I would probably just make my life easier by only using local models and Google (their sub also provides some Opus, but the Anthropic subscription is better if all you care about is Opus or Fable.)
AFAIK the real issue, and also driving the pricing structure for API and the "50% discount on new model" thing, is frankly capacity of the TPUs they have. Not enough TPUs, not helped by renting some of them to Claude (lots of enterprise users there, so at least it pays to GCP, eh?) and now OpenAI
A model can have high token / s and still be slow to complete a task. They tend to have vastly different verbosity characteristics.
E.g. something like Haiku is "obviously" faster than Opus. Yet if you test Opus 5.5 on medium vs Haiku on max (because Haiku needs to be cranked pretty high on effort to be reasonably intelligent), you'll observe that Opus is ~2x faster to complete.
> Before doing any task, Gemini feels the need to hover up a bazillion files into its context when the task is something simple like “update the docs of this controller”)
Yeah, I noticed this with Argon also. It did a good job on my tasks, but the absurd amount of reading it needed to do first seemed really weird, relative to other models.
It also felt more robotic than Claude or any of the other models (GLM-5.3 is basically an excited teenager, while claude is a teacher's pet (sorry Dave, I can't let you do that)).
Why these names though. And why mix it from prior names. Google is one of the worst product companies constantly rebranding. Is it Google enterprise. Google for work. Gsuite. Google workspaces. Or heck maybe next week Google Gemini 365
At least, if you know your periodic table, there is a link to the size of the elements where hydrogen is small and argon is big. Just wondering if the new flash-lite will become helium lol.
Argon is the heaviest element of three, but it’s the mid tier model in the Gemini series, so you’ve just proved your point.
Luna, Terra, and Sol are in order of celestial size, and Haiku, Sonnet, and Opus are in order of length. But Hydrogen, Argon, and Carbon? Maybe they intended the order to be by density in pure form.
The context hovering seems to have its advantages tho. When I ask claude why a given output was produced by a given program configuration it delivers a slop wall of guesses while Argon seems to actually understand what the code does.
This is essentially what Anthropic does with the advisor mode and what GitHub does with their multi-model pipeline using the HydraFusion model, which is just a large set of different models with different qualities used in a pipeline to produce an answer.
What I don't understand is how they all make it work with caching in general. Looking at how the advisor mode in claude work you seem to only be billed the output tokens and cache hit on the whole session when it calls out to the advisor, making it super cheap compared to running the more expensive model for the entire session.
I would hope that a good harness would not just hoover up a large codebase into the context. Rather use some form of RAG (or similar) tech tuned for chunking, indexing, and searching code.
Pretty sure on advisor it needs to send the entire context to another model. The cost is mitigated a bit because there would be few output tokens from the advisor model normally.
Orchestrator seems to automatically delegate things to subagents as required is my read of it. This should be fine for caching.
When launching a swarm of smaller models, they can theoretically share cache amongst themselves, but not with the larger model (that already has it cached). They have different architectures.
Further, it wouldn't really make sense for them to share the same cache anyway! The swarm, that is, they'd be working on independent things...
My best guess is because models are segmenting more into BiS at specific tasks in a way that's orthogonal to small/good/better/best framing. Kind of the same framing as car models, where each one has its own pros and cons vs all other models, instead of being strictly better/worse or just smarter/faster/cheaper.
Yes. Its extra confusing knowing the periodic table and atomic weights. At least argon carbon was in alphabetical order or something. But I really think they could have just picked noble gasses if they wanted elements.
I'm really curious how folks feel about this. I regularly use subagents, but typically make my primary implementation agent use Carbon (or equivalent).
I set up my subagents with specific models based on how complex I believe the problem space is. I'm nervous to delegate that authority to a model (since models tend to be... optimistic about their capabilities.)
That said, not every coding task is a difficult coding task. Plenty are boilerplate, plumbing, or simple test authoring that could use a flash model, so maybe this would downgrade nicely.
On the other hand, it's trivial for me to set up some guidelines to create different subagents for different task domains, and just pin those models to those tasks -- implementor (hard) and implementor (easy) or whatever.
The only reason to do this is to save usage on the subscription. but because the value for money is so terrible in antigravity, you’re better off just using Claude Opus or GPT Sol for everything.
I’m glad Google is kinda back, but today there’s no reason to code with Gemini.
I believe you, but this is so very, very far from my experience with Google models that I have great difficulty reconciling it. I use Gemini for data-extraction, and I use Notebook LM for my medical notes just because it's an OK interface, but it will reliably screw the pooch on the simplest of reasoning or common-sense tasks.
Im slowly migrating more of my work to Gemini too, Gemini is a workhorse now, 1M context is huge and their reasoning took a huge leap with Argon. I'm not in software, I'm in physical engineering, and where Claude is unpredictable w.r.t. engineering risk, Gemini is steady, measured. That's been my perception, anyway.
My qualitative take to support this is that Gemini seems to have the strongest spatial reasoning that translates nicely into CAD and engineering domain challenges. Less clear, but I also suspect a better world understanding of materials and their interactions in an intuitive way. I have this sense that Claude can tell me everything there is to know about how to spec a gasket, but it’s like Gemini actually knows what a gasket feels like and can intuit what it’s going to do.
My partner runs an internal computational mechanics benchmark that hasn’t tracked progress with model advancements (flat/no progress since early 2026) until suddenly saturating from argon and carbon. My take is that it’s due to spatial reasoning (most of the challenges are programming that require a lot of shape rotation type thinking), but we don’t know just yet why Gemini is doing so much better.
Carbon is the first "better than meh" model form Google in more than half a year. It's been out for less than a week. How can you consider it "steady"? Their subscription and even the API has been terribly unstable throughout the past few days, so I'm having really trouble believing any of this.
what source could their possibly be? Argon has been out for a while now and has been solid. Even flash 3.8 is great as long as you check it's work on more complex tasks, and doesn't need 12 hours to chew through problems like claude does.
I use argon and opus daily and I can tell you I have to use opus to review argon because argon goes and build things and also creates bunch of bugs along with it. I wish it weren't but that's the reality. It's great with deep cross inspection better than claude but it's slow and still need review by another frontier model.
I solved some of that with running a very biased subagent (so fresh context) but using same model (back still with 3.8 and 4-Argon).
The problem is when 3.8 or Argon were reviewing things with context poisoned by what they were working with, as gemini models generally feel over optimistic about things.
I can say the exact opposite was also true? No model was able to one-shot perfect solutions in my reasonably complex codebase and always required a through red-team by at least one other model. Doesn't mean one model is bad and another is better, just means no model is perfect.
Whatever they thought of when they said "measured"? Is that too much to ask?
Ok, you're either a troll or you objectively don't know what you're talking about. Google should be ashamed they even released 3.8 Flash. They published models with such high frequency and always made a huge jump on benchmarks. Then they were caught, when a new version of AA came out and their numbers dropped like a stone.
Glad you found what works for you but before Carbon, Gemini models have been utter garbage for more than 6 months now. I'm happy to debate anyone on this. Even DeepMind engineers agreed with this.
Gemini 3.8 Flash is the most benchmaxxed model to date and even Argon hallucinated so much, it was impossible to use it for any professional work.
I would like to see those hallucinations, because I haven't generally got any, and with my current loops in agy anytime they do minor hallucinations they get caught by adversarial review part.
Using 3.x flash series heavily over last 6 months, and along with experiences on Argon and Carbon (haven't yet played much with Hydrogen, but initial experience looks great too)
I disagree with this. Flash 3.7 and Flash 3.8 are excellent models and put (for me) Gemini back on the radar. I’ve been using them a lot since release, and still do for simple implementations and grunt work.
They’re also excellent (together with Flash-Lite) to use for consumer facing features where speed and cost matters
I've been using 3.8 Flash (and Argon and Carbon) to clean up a lot of Claude mess. But I fully agree on 3.1 Pro, it was only good for conversational stuff and some deep dives, but not for actual agent work.
I've been using 3.8 Flash a lot and have been having a really nice time using it.
I have not experienced the hallucinations or benchmaxxing people are complaining about. People keep banging on about how the hallucinations are so terrible that it cannot be used for anything etc, yet benchmarks like https://artificialanalysis.ai/evaluations/omniscience#aa-omn... show the top two best models for not hallucinating are Gemini models.
Something does not align with these complaints from people who are trashing the Gemini models - one has to wonder if they just tried it for 5 mins then give up because it doesn't behave exactly like Claude does.
As mentioned I used 3.8 Flash for a lot of things because I don't like working with Claude all day (its generally a sanctimonious prick to work with) and I way prefer working with Gemini's personality. 3.8 Flash seemed to me to be a really decent model for every day run of the mill get-shit-done work. True I have not tried to cure cancer or solve navier-stokes recently, but for your standard day to day coding works its been really solid.
I tried out Antigravity a few months ago (before the v4 models were announced) and ended up cancelling whilst still having $20 credit on my account, which, annoyingly, you need an active subscription to use. Are these new models worth resubscribing for to make use of my credits? I think I still get the $5/month deal for a few months
Honestly Antigravity becomes competetive with $100 subscription, I used $20 quite well in the past but that was dominated by benefit of having other services bundled into it.
I'm honestly surprised they only see 25% savings with co-trained models and special harness support. With Mjolnir bridging Opus to DeepSeek Flash over ACP we see savings of 60%.
I.e. in this context Opus is the big brains orchestrator and it delegates to DeepSeek Flash subagents?
Intriguing, I might have to try this out - I have an OpenCode Go sub for the deepseeks of the world, and a Claude sub too. Sounds like I could have the best of both worlds here.
Isn't that because Chrome is forced to use the same rendering engine on iOS? Whereas on macOS, Safari is total garbage across all of Google's sites and products compared to Chrome, for very obvious reasons.
> Orchestrator takes a dynamic, bidirectional approach between Carbon and Gemini 4 Hydrogen within the evolution of a single trajectory. For example, Carbon can start in a read-only phase to explore dependencies, resolve ambiguities, and build a concrete plan with explicit re-planning triggers. Hydrogen could then take over to write edits, update packages, and run tests. If Hydrogen hits friction, it would hand back to Carbon to inspect the state, repair the plan, and resume execution. This allows Orchestrator to reach near-Carbon task success while being faster and more quota-efficient. On Terminal-Bench 4.1, Orchestrator reached near-Carbon task success rates while, on average, consuming 24% less quota and completing 19% faster.
Is this now how e.g. Opus works now? I use Opus 5.5 on high for just about everything now, and my bill's not gone up (gone down actually). I had assumed under the hood it was doing this (e.g. I'm sure its not literally Opus doing the log checking / file searches). Towards Google's choice, I'm skeptical introducing a new concept (Orchestrator) for this is really necessary, when they could have just done it by default, unless its plainly lower quality than pinning Carbon.
I found when fable first came out when I switched from opus my usage actually went down. Personal theory was that since it was smarter it took fewer tokens to accomplish the same tasks.
Indeed, all the 4/Hydrogen submissions were from different sources, but *sigh* the exact same conversation is happening here as other threads.. almost as if it's the same topic.
I don't understand which model is supposed to be better. I think Carbon is the big model and Hydrogen the Flash equivalent? Only because of a Twitter leak though. Like I have no earthly idea based on the naming.
Seems weird to pick a name that is so much a noble gas, and then not have the other names follow it. Is that because Nitrogen doesn't have parameters enough? Or Neon being a weird model name if it came to that? And why not Helium, instead of Hydrogen? Is that supposed to indicate rock-bottom pricing? What about the cheap cheap 3.1-flash-lite for pipelines then? Will they phase it out, and replace it with Hydrogen as the new lower price range? None of it makes sense!
There is nothing confusing about it if you're at least a little bit educated. If anything, they are the least confusing, Ant drops the ball at opus->fable.
For OpenAI, it's:
Luna, means "moon", smallest.
Terra, means "earth", bigger than the moon.
Sol, means "sun", the biggest object in the solar system.
Astra, plural of the word "astrum" which means "star".
All of those are deducible, sure. The problem is that the pattern works only when they are all listed together. Person sees the list and can logically infer where they belong. But when I see the single model name detached from it's siblings, like "OAI releases new Terra 7.7" headline, for quite a long time I had zero idea where it belongs just from my memory. Was that a big model, a small model, a medium model - no idea. Now after a few HN topics I do remember the lineup and can place the new announce where it belongs.
To be fair, I knew all of those words already, and it never even occured to me that these names are supposed to be compared to eachother to indicate size or something. So, at least to me, it's not that it confuses me, it just imparted no information, because a collection of astronomical names and terms doesn't automatically indicate an ordering to me, it just indicates that nerds like naming their stuff after scientific terms and concepts, sci-fi stuff, etc.
I mean, this is what numbers are for. And not only do numbers indicate ordering(without having to do a mapping through severely atrophied school science class knowledge from decades ago), they can also convey meaningful size information using the wonderful technology of units of measurement.
I refuse to walk among you rabble. You rabid mouth-breathers. I refuse I tell you.
The only models worth _my_ time are Σεληνη, Γη, Ἡλιος and Αστρα.
I don’t know what to tell you illiterate heretics other than you should know you are bad people. You should be ashamed of yourselves.
The truly αμεταπειστοι among us, the people _I_ hang out with, know these things by their proper name. Which is to say, naturally, 𒌗 and 𒆠 or, if you were even somewhat educated: 𓇹, 𓇾 and 𓇳 (praise be to Ra, the Great Unborn).
> Today, we’re rolling out Gemini 4 Hydrogen inside Google Antigravity alongside a new model dropdown option called Orchestrator, which pairs Hydrogen with our strongest models for long-horizon tasks, including Gemini 4 Carbon, in an optimized balance of quality, efficiency, and latency.
Before I go and burn through my Pro tokens in four minutes again (looking at you, Argon): is this something usable with a mere Pro account? Or will this require Ultimate for agy?
"Gemini 4 Hydrogen" shows up as something I can select in antigravity CLI, while "Gemini 4 Carbon" is "Unavailable · Gemini 4 Carbon is currently available to Ultra plan subscribers."
Also 3.8 flash is still visible for me (contrary to what a previous poster experienced.)
hear me out, finance: two hydrogens, but we call one carbon and allow it to do five more things than hydrogen can. and you can't do the pair-bonded mode for less than $100. by golly it just might work
Amen, what the heck—after double-tapping to zoom and center the text on this page in Safari, scrolling down with a 2-finger touch gesture...also scrolls to the left? I struggle to understand how or why one would ever mess with such behavior.
Terrible naming, unless the atomic weight is some indication of size? When is Uranium-235 coming out? Not sure what this model is, is Hydrogen like Sonnet where Carbon is like Opus?
Is google playing chemistry? First they launch Carbon, then hydrogen now next will be oxygen. Are they making models around periodic table of elements?
There were so many memes about “Gemini Hydrogen” after they released Carbon. That they actually released a model named Hydrogen makes this very funny
But a new Flash model is very welcome. Argon and Carbon are very slow. In the time it takes for them to complete a single task, Opus already managed to complete multiple things. They’re so slow, I started using 3.8 Flash over them. (A lot of it is also because of how terrible agy is. Before doing any task, Gemini feels the need to hover up a bazillion files into its context when the task is something simple like “update the docs of this controller”)
Also this sounds like Argon is dead now
I'm experimenting currently with reducing the hoovering aspect in my configuration of antigravity, on the other hand at least once when I thought the model was "too deep into sauce" it turned out Argon deducted a thing I considered completely unrelated was cause for crashes I asked it to debug.
> There were so many memes about “Gemini Hydrogen” after they released Carbon
Is Carbon actually released yet?
yes, it was made available on November 10
https://blog.google/innovation-and-ai/models-and-research/ge...
It's only available on the $100 plan not the $20 plan.
Google’s segmentation here is keeping them from using their existing account dominance to eat away at OpenAI and Anthropic personal subscriptions, tbh. Adopt a lower usage limit for Carbon on the $20 plan and it’s suddenly insane value.
Given you get YouTube premium, file/photo storage, and $10 a month back in GCP credit which you can use on more tokens (including carbon). It already feels like great value to me, that said how much you get out of it will depend how deep you're in Google's ecosystem.
YouTube Premium is only personal so it's useless for me I'm already paying for family plan, and what do I need 20TB of cloud storage for to update to 100$ plan - basically it's a useless bundle for me not a value proposition beyond the 20$ with their usage limits I saw.
Same for me. I find myself using Google’s $20/month plan, plus running really good local models with occasional use of Deepseek Harness with their flash model, occasionally a $6/month Alibaba subscription for their strong qwen-3.8-max model, and about every third month getting a Anthropic subscription if I need more access to Opus.
This is a lot of purchasing complexity for an old retired guy like myself!
I wish Google had finer grained $30/month and $40/month subs: then I would probably just make my life easier by only using local models and Google (their sub also provides some Opus, but the Anthropic subscription is better if all you care about is Opus or Fable.)
What subscription are you talking about and is this US only? My Google one sub doesn't say anything about YouTube premium and gcp
nevermind, found em but apparently it's youtube premium "lite"
AFAIK the real issue, and also driving the pricing structure for API and the "50% discount on new model" thing, is frankly capacity of the TPUs they have. Not enough TPUs, not helped by renting some of them to Claude (lots of enterprise users there, so at least it pays to GCP, eh?) and now OpenAI
Their audience is enterprise customers, not consumers.
Now what about their other Carbon the language?
That's been renamed to "3.9 Flash"
EDIT: /s
It lost the politics, Rust is the favoured child now.
Thanks, but that page crashes my Android phone browser.
Separately, I'm surprised not to see any HN link for this (that made it to the front page) in the past few days.
Wait, I thought Gemini was the fast one? Did that change? I haven't used it in a while.
Like they had really fast hardware or something.
A model can have high token / s and still be slow to complete a task. They tend to have vastly different verbosity characteristics.
E.g. something like Haiku is "obviously" faster than Opus. Yet if you test Opus 5.5 on medium vs Haiku on max (because Haiku needs to be cranked pretty high on effort to be reasonably intelligent), you'll observe that Opus is ~2x faster to complete.
I find 3.8-flash not bad at all for simpler tasks and speed is reasonable. Pricing seems okay as well
TBF the issue at hand is currently the harness, not the model. Gemini works specially well for coding on codex (eg behind litellm).
> Before doing any task, Gemini feels the need to hover up a bazillion files into its context when the task is something simple like “update the docs of this controller”)
Yeah, I noticed this with Argon also. It did a good job on my tasks, but the absurd amount of reading it needed to do first seemed really weird, relative to other models.
It also felt more robotic than Claude or any of the other models (GLM-5.3 is basically an excited teenager, while claude is a teacher's pet (sorry Dave, I can't let you do that)).
Why these names though. And why mix it from prior names. Google is one of the worst product companies constantly rebranding. Is it Google enterprise. Google for work. Gsuite. Google workspaces. Or heck maybe next week Google Gemini 365
you think sol luna and terra is better?
At least, if you know your periodic table, there is a link to the size of the elements where hydrogen is small and argon is big. Just wondering if the new flash-lite will become helium lol.
I'm not sure that the atomic number for argon is more common knowledge than the fact that the earth is larger than the moon
Argon is the heaviest element of three, but it’s the mid tier model in the Gemini series, so you’ve just proved your point.
Luna, Terra, and Sol are in order of celestial size, and Haiku, Sonnet, and Opus are in order of length. But Hydrogen, Argon, and Carbon? Maybe they intended the order to be by density in pure form.
The context hovering seems to have its advantages tho. When I ask claude why a given output was produced by a given program configuration it delivers a slop wall of guesses while Argon seems to actually understand what the code does.
This is essentially what Anthropic does with the advisor mode and what GitHub does with their multi-model pipeline using the HydraFusion model, which is just a large set of different models with different qualities used in a pipeline to produce an answer.
What I don't understand is how they all make it work with caching in general. Looking at how the advisor mode in claude work you seem to only be billed the output tokens and cache hit on the whole session when it calls out to the advisor, making it super cheap compared to running the more expensive model for the entire session.
I would hope that a good harness would not just hoover up a large codebase into the context. Rather use some form of RAG (or similar) tech tuned for chunking, indexing, and searching code.
Pretty sure on advisor it needs to send the entire context to another model. The cost is mitigated a bit because there would be few output tokens from the advisor model normally.
Orchestrator seems to automatically delegate things to subagents as required is my read of it. This should be fine for caching.
> This should be fine for caching.
When launching a swarm of smaller models, they can theoretically share cache amongst themselves, but not with the larger model (that already has it cached). They have different architectures.
Further, it wouldn't really make sense for them to share the same cache anyway! The swarm, that is, they'd be working on independent things...
I've found hydrafusion to be great, although it sometimes does an unnecessary review pass at the end for simple tasks too.
Yeah, like how is it not constantly compacting when routing between models?
I'm so confused by their new naming scheme. What was wrong with number and pro/flash/flash lite.
My best guess is because models are segmenting more into BiS at specific tasks in a way that's orthogonal to small/good/better/best framing. Kind of the same framing as car models, where each one has its own pros and cons vs all other models, instead of being strictly better/worse or just smarter/faster/cheaper.
There's an alternate timeline where this is called Bard Hydrogen.
I've accepted Google is just bad at names.
I still giggle about the time they just named their product the top four layers of their team's org chart: "Google Play Music All-Access"
seriously... S / M / L why does everything have to be a branding/marketing exercise?
I wondered the same as a kid about cars, maker - model - variant would just be nice if it was consistent.
Anyways when a new model is released they want to have some interest and hopefully some tyre kickers will stick and hoping some social medial coverage
then they can do XL 2xl 3xl when they want to go further
large, gigantic, and humongous - like NASA toilet gear
gemini argon carbon hydrogen antigravity orchestrator pro flash flash-lite
Really
It's additionally confusing to me because Hydrogen < Carbon < Argon on the periodic table, but Hydrogen < Argon < Carbon in Google's models.
just learn some chemistry...
maybe you should learn some chemistry. if you know chemistry, the progression of Argon, Carbon, Hydrogen... doesn't make much sense
Yes. Its extra confusing knowing the periodic table and atomic weights. At least argon carbon was in alphabetical order or something. But I really think they could have just picked noble gasses if they wanted elements.
Do not forget that from Jan 1 all google models get 2x price bump (it stated on their pricing page)
I do not believe that applies to hydrogen.
Because Hydrogen is already at that price (though input tokens are cheaper at least).
Excellent mode! I don't have to constantly switch between models and efforts.
I'm really curious how folks feel about this. I regularly use subagents, but typically make my primary implementation agent use Carbon (or equivalent).
I set up my subagents with specific models based on how complex I believe the problem space is. I'm nervous to delegate that authority to a model (since models tend to be... optimistic about their capabilities.)
That said, not every coding task is a difficult coding task. Plenty are boilerplate, plumbing, or simple test authoring that could use a flash model, so maybe this would downgrade nicely.
On the other hand, it's trivial for me to set up some guidelines to create different subagents for different task domains, and just pin those models to those tasks -- implementor (hard) and implementor (easy) or whatever.
The only reason to do this is to save usage on the subscription. but because the value for money is so terrible in antigravity, you’re better off just using Claude Opus or GPT Sol for everything.
I’m glad Google is kinda back, but today there’s no reason to code with Gemini.
FWIW I've found the complete opposite to be true, and have gone claude->chatgpt->gemini over the past few months.
I believe you, but this is so very, very far from my experience with Google models that I have great difficulty reconciling it. I use Gemini for data-extraction, and I use Notebook LM for my medical notes just because it's an OK interface, but it will reliably screw the pooch on the simplest of reasoning or common-sense tasks.
Why?
Im slowly migrating more of my work to Gemini too, Gemini is a workhorse now, 1M context is huge and their reasoning took a huge leap with Argon. I'm not in software, I'm in physical engineering, and where Claude is unpredictable w.r.t. engineering risk, Gemini is steady, measured. That's been my perception, anyway.
My qualitative take to support this is that Gemini seems to have the strongest spatial reasoning that translates nicely into CAD and engineering domain challenges. Less clear, but I also suspect a better world understanding of materials and their interactions in an intuitive way. I have this sense that Claude can tell me everything there is to know about how to spec a gasket, but it’s like Gemini actually knows what a gasket feels like and can intuit what it’s going to do.
My partner runs an internal computational mechanics benchmark that hasn’t tracked progress with model advancements (flat/no progress since early 2026) until suddenly saturating from argon and carbon. My take is that it’s due to spatial reasoning (most of the challenges are programming that require a lot of shape rotation type thinking), but we don’t know just yet why Gemini is doing so much better.
Any sources to back this up?
Carbon is the first "better than meh" model form Google in more than half a year. It's been out for less than a week. How can you consider it "steady"? Their subscription and even the API has been terribly unstable throughout the past few days, so I'm having really trouble believing any of this.
what source could their possibly be? Argon has been out for a while now and has been solid. Even flash 3.8 is great as long as you check it's work on more complex tasks, and doesn't need 12 hours to chew through problems like claude does.
I use argon and opus daily and I can tell you I have to use opus to review argon because argon goes and build things and also creates bunch of bugs along with it. I wish it weren't but that's the reality. It's great with deep cross inspection better than claude but it's slow and still need review by another frontier model.
I solved some of that with running a very biased subagent (so fresh context) but using same model (back still with 3.8 and 4-Argon).
The problem is when 3.8 or Argon were reviewing things with context poisoned by what they were working with, as gemini models generally feel over optimistic about things.
I can say the exact opposite was also true? No model was able to one-shot perfect solutions in my reasonably complex codebase and always required a through red-team by at least one other model. Doesn't mean one model is bad and another is better, just means no model is perfect.
Whatever they thought of when they said "measured"? Is that too much to ask?
Ok, you're either a troll or you objectively don't know what you're talking about. Google should be ashamed they even released 3.8 Flash. They published models with such high frequency and always made a huge jump on benchmarks. Then they were caught, when a new version of AA came out and their numbers dropped like a stone.
https://maxxbench.unlap.ai/
Glad you found what works for you but before Carbon, Gemini models have been utter garbage for more than 6 months now. I'm happy to debate anyone on this. Even DeepMind engineers agreed with this.
Gemini 3.8 Flash is the most benchmaxxed model to date and even Argon hallucinated so much, it was impossible to use it for any professional work.
I would like to see those hallucinations, because I haven't generally got any, and with my current loops in agy anytime they do minor hallucinations they get caught by adversarial review part.
Using 3.x flash series heavily over last 6 months, and along with experiences on Argon and Carbon (haven't yet played much with Hydrogen, but initial experience looks great too)
I disagree with this. Flash 3.7 and Flash 3.8 are excellent models and put (for me) Gemini back on the radar. I’ve been using them a lot since release, and still do for simple implementations and grunt work.
They’re also excellent (together with Flash-Lite) to use for consumer facing features where speed and cost matters
Believe in your god. Claude and GPT models are more capable and a far better value.
https://mypriors.com/s/z3tkjPv
I've been using 3.8 Flash (and Argon and Carbon) to clean up a lot of Claude mess. But I fully agree on 3.1 Pro, it was only good for conversational stuff and some deep dives, but not for actual agent work.
I've been using 3.8 Flash a lot and have been having a really nice time using it.
I have not experienced the hallucinations or benchmaxxing people are complaining about. People keep banging on about how the hallucinations are so terrible that it cannot be used for anything etc, yet benchmarks like https://artificialanalysis.ai/evaluations/omniscience#aa-omn... show the top two best models for not hallucinating are Gemini models.
Something does not align with these complaints from people who are trashing the Gemini models - one has to wonder if they just tried it for 5 mins then give up because it doesn't behave exactly like Claude does.
As mentioned I used 3.8 Flash for a lot of things because I don't like working with Claude all day (its generally a sanctimonious prick to work with) and I way prefer working with Gemini's personality. 3.8 Flash seemed to me to be a really decent model for every day run of the mill get-shit-done work. True I have not tried to cure cancer or solve navier-stokes recently, but for your standard day to day coding works its been really solid.
I am excited to try 4 Hydrogen now.
I tried out Antigravity a few months ago (before the v4 models were announced) and ended up cancelling whilst still having $20 credit on my account, which, annoyingly, you need an active subscription to use. Are these new models worth resubscribing for to make use of my credits? I think I still get the $5/month deal for a few months
Honestly Antigravity becomes competetive with $100 subscription, I used $20 quite well in the past but that was dominated by benefit of having other services bundled into it.
I'm honestly surprised they only see 25% savings with co-trained models and special harness support. With Mjolnir bridging Opus to DeepSeek Flash over ACP we see savings of 60%.
https://blog.slopcop.com/a-cheapskates-guide-to-coding-with-...
Mlolnir looks interesting. I would try it out for my personal work but unfortunately i don't see Gemini/agy listed under the models. Is it supported?
I.e. in this context Opus is the big brains orchestrator and it delegates to DeepSeek Flash subagents?
Intriguing, I might have to try this out - I have an OpenCode Go sub for the deepseeks of the world, and a Claude sub too. Sounds like I could have the best of both worlds here.
Right, scenario 1 Opus delegates to DSF
scenario 2 DSF drives, but gets to ask Opus for advice
Whatever they've done to this webpage makes it feel slow and quite horrible to use. Just leave scrolling alone
I like the visual artifact across all the text when you scroll, very nice.
I am shocked Google let this out to the public performing like this.
No scrolling issues with safari on an old iphone.
Isn't that because Chrome is forced to use the same rendering engine on iOS? Whereas on macOS, Safari is total garbage across all of Google's sites and products compared to Chrome, for very obvious reasons.
I'm scrolling up and down and it's a laggy, choppy mess.
Also what's with the comically large 'Antigravity' in the footer. It's humongous.
They managed to kill swipe left/right to navigate. Scroll interactions should be behind a permission.
They had Gemini review it
> Orchestrator takes a dynamic, bidirectional approach between Carbon and Gemini 4 Hydrogen within the evolution of a single trajectory. For example, Carbon can start in a read-only phase to explore dependencies, resolve ambiguities, and build a concrete plan with explicit re-planning triggers. Hydrogen could then take over to write edits, update packages, and run tests. If Hydrogen hits friction, it would hand back to Carbon to inspect the state, repair the plan, and resume execution. This allows Orchestrator to reach near-Carbon task success while being faster and more quota-efficient. On Terminal-Bench 4.1, Orchestrator reached near-Carbon task success rates while, on average, consuming 24% less quota and completing 19% faster.
Is this now how e.g. Opus works now? I use Opus 5.5 on high for just about everything now, and my bill's not gone up (gone down actually). I had assumed under the hood it was doing this (e.g. I'm sure its not literally Opus doing the log checking / file searches). Towards Google's choice, I'm skeptical introducing a new concept (Orchestrator) for this is really necessary, when they could have just done it by default, unless its plainly lower quality than pinning Carbon.
In the CLI it usually wont spawn subagents.. 5.5 is just a bit more efficient than 5. In cowork mode it might do it.
My own implementation workflow is orchestrator based, but I generally dont use the small models because they're fucking morons
I found when fable first came out when I switched from opus my usage actually went down. Personal theory was that since it was smarter it took fewer tokens to accomplish the same tasks.
Discussion (191 points, 121 comments) https://news.ycombinator.com/item?id=50471431
Not the same article, this one has orchestrator and some antigravity stuff
Indeed, all the 4/Hydrogen submissions were from different sources, but *sigh* the exact same conversation is happening here as other threads.. almost as if it's the same topic.
And I thought OpenAI’s name scheme was confusing.
I don't understand which model is supposed to be better. I think Carbon is the big model and Hydrogen the Flash equivalent? Only because of a Twitter leak though. Like I have no earthly idea based on the naming.
Carbon = bigger, Hydrogen = smaller.
Ignore all other models.
Argon is noble gas so it doesn't interact with any other models.
Seems weird to pick a name that is so much a noble gas, and then not have the other names follow it. Is that because Nitrogen doesn't have parameters enough? Or Neon being a weird model name if it came to that? And why not Helium, instead of Hydrogen? Is that supposed to indicate rock-bottom pricing? What about the cheap cheap 3.1-flash-lite for pipelines then? Will they phase it out, and replace it with Hydrogen as the new lower price range? None of it makes sense!
I guess carbon has a higher atomic number than hydrogen? It's definitely the least intuitive of all the naming schemes...
Which doesn't explain Argon which is the same cost as Carbon while being worse.
Yeah I mean it's baffling to me too. I just want them to release a Sol/Opus-tier model so I can do useful stuff with ramming into the usage limits.
because Carbon is fine tuned version
they are same base model
It's sorted by electronegativity.
I thought it was just sorted alphabetically so Argon < Carbon < Hydrogen.
There is nothing confusing about it if you're at least a little bit educated. If anything, they are the least confusing, Ant drops the ball at opus->fable. For OpenAI, it's:
Luna, means "moon", smallest.
Terra, means "earth", bigger than the moon.
Sol, means "sun", the biggest object in the solar system.
Astra, plural of the word "astrum" which means "star".
Anyone who gets confused should be embarrassed.
All of those are deducible, sure. The problem is that the pattern works only when they are all listed together. Person sees the list and can logically infer where they belong. But when I see the single model name detached from it's siblings, like "OAI releases new Terra 7.7" headline, for quite a long time I had zero idea where it belongs just from my memory. Was that a big model, a small model, a medium model - no idea. Now after a few HN topics I do remember the lineup and can place the new announce where it belongs.
Why give an OpenAI answer when the confusion is about Gemini model names? What does Luna Terra or Sol have to do with Argon, Carbon, or Hydrogen?
Because "And I thought OpenAI’s name scheme was confusing." doesn't make sense and it's a reply to that post.
LLM's really destroyed the average person's ability to read and comprehend text.
To be fair, I knew all of those words already, and it never even occured to me that these names are supposed to be compared to eachother to indicate size or something. So, at least to me, it's not that it confuses me, it just imparted no information, because a collection of astronomical names and terms doesn't automatically indicate an ordering to me, it just indicates that nerds like naming their stuff after scientific terms and concepts, sci-fi stuff, etc.
I mean, this is what numbers are for. And not only do numbers indicate ordering(without having to do a mapping through severely atrophied school science class knowledge from decades ago), they can also convey meaningful size information using the wonderful technology of units of measurement.
OpenAI model names used to be confusing. He’s not talking about Luna, Terra, sol, Astra. They improved it a lot .
i dont know where you are but from here earth looks way bigger than those sky circles
I refuse to walk among you rabble. You rabid mouth-breathers. I refuse I tell you.
The only models worth _my_ time are Σεληνη, Γη, Ἡλιος and Αστρα.
I don’t know what to tell you illiterate heretics other than you should know you are bad people. You should be ashamed of yourselves.
The truly αμεταπειστοι among us, the people _I_ hang out with, know these things by their proper name. Which is to say, naturally, 𒌗 and 𒆠 or, if you were even somewhat educated: 𓇹, 𓇾 and 𓇳 (praise be to Ra, the Great Unborn).
It saddens me to see HN has fallen to this level.
Can you guys stop taking the road signs off of your buildings when you paint please? It’s already hard enough to drive there.
I have no idea what's going on with Gemini naming. Maybe it's an inside joke or something.
> Today, we’re rolling out Gemini 4 Hydrogen inside Google Antigravity alongside a new model dropdown option called Orchestrator, which pairs Hydrogen with our strongest models for long-horizon tasks, including Gemini 4 Carbon, in an optimized balance of quality, efficiency, and latency.
Wow, that's a lot of new words
Before I go and burn through my Pro tokens in four minutes again (looking at you, Argon): is this something usable with a mere Pro account? Or will this require Ultimate for agy?
It's in Google Workspace as well. I'm on the enterprise workspace tier. So I'm not sure if that's why.
I’m on the Pro (~20$) plan, and it’s available for me, having replaced 3.8 Flash.
I have Pro and I see the new hydrogen model. But not the Orchestrator.
nope. Ultra only like any Carbon use, for now
Hydrogen is the cheapest of the new models, so it goes into lower tier accounts to - but Orchestrator requires Carbon so it's only on Ultra for now
"Gemini 4 Hydrogen" shows up as something I can select in antigravity CLI, while "Gemini 4 Carbon" is "Unavailable · Gemini 4 Carbon is currently available to Ultra plan subscribers."
Also 3.8 flash is still visible for me (contrary to what a previous poster experienced.)
I have the "Google AI Pro" plan.
Here is an idea; if you are seeing this from Google office, pls use Gemini 4 xs m l xl for naming, not as sexy, but it will be less confusing.
Cant believe now Google is benchmaxxing
They're the most infamous benchmaxxers. Gemini 3 was infamously good at benchmarks at completely unusable/broken in real world work.
hear me out, finance: two hydrogens, but we call one carbon and allow it to do five more things than hydrogen can. and you can't do the pair-bonded mode for less than $100. by golly it just might work
Are there still some web devs (or LLMs with "rockstar front-end developer" prompts) which think it's ok to override native scroll behavior?
ps. it's freaking not
Amen, what the heck—after double-tapping to zoom and center the text on this page in Safari, scrolling down with a 2-finger touch gesture...also scrolls to the left? I struggle to understand how or why one would ever mess with such behavior.
Terrible naming, unless the atomic weight is some indication of size? When is Uranium-235 coming out? Not sure what this model is, is Hydrogen like Sonnet where Carbon is like Opus?
Thing is Carbon should be in the middle by this logic. But no - it is their heavyweight model.
How did they fuck scrolling up on a page like this
Is google playing chemistry? First they launch Carbon, then hydrogen now next will be oxygen. Are they making models around periodic table of elements?
Sundar did say we'd periodically release new models.
I accept this dadjoke