Before adding more models, can we please fix the ridiculous quota system?
Every time a new “powerful” model gets added, it ends up being so heavily restricted that a few prompts are enough to burn through the entire session allowance. Sometimes even simple, lightweight tasks can trigger the limit, and then you’re stuck waiting 24 hours—or even days—for a refresh.
What’s the point of adding more models if users can barely use the ones already available?
Before adding models you must you have to understand one thing that Google has no access to anthropic models free of cost even they host them and fully maintain them with their own budget and money the reason is that even they make their own load balancer as their own servers to handle the load and host the model but they still have to pay the money or the payment as per the deal whatever they have with anthropic with the model define by class size terminology and price negotiated so for each query and token your spend even they cost Google at their hardware level if they host the model but they also cost Google as a software term negotiation level let they have to pay anthropic the amount per usage of the model FB are talking about offers or the fable class of models definitely the price in will be two high for Google to pay for anthropic which is not a good deal for Google like it’s like just spending your own money to raise your competitor in your own environment that’s the reason why Google forces you to use Gemini which has even higher Kota but refreshes in 5 hours where as if you see other model providers you barely get some usage even in the pro or ultra accounts because the reason is simple for query and for the usage Google is paying twice one for the hardware cost and the load balancer cost and the second for the software cost that it needs to pay to the origin company for using their model in their own environment if you desperately need that you can just simply by anthropic API credits or simply sign up with their claude Pro or max plan which will definitely give you more value as compared to Google anti gravity it will be a wastage of time and wastage of efforts to wait for the model to be released in anti gravity as its not possible even if they somehow add on negotiate the price you will barely get the usage like how you get the Gemini one in the IDe it’s not like that Google has unlimited money bank to spend on they will definitely priorities their own models and they will not pay for other competitors for the same environment they own answer is simple don’t wait by the official one use claude code
I understand the point some people are making in the comments/replies.
However, as the Antigravity supports top ups and monthly AI credits, I suppose it is doable, just like how claude opus 4.6 is there and other couple of models (sonnet and GPT OSS).
This being said, it is understandable that Google prefers their users to use their own models, but Models are slowly (or quickly depending on your perspective of time) becoming a commodity… USP of antigravity is their dual nature - Codex and Claude like interface for simple conversational vibe coding + IDE like Cursor for advanced stuff…
If google leverages IDE and extension store kind of thing… it can be really useful for specific tasks… It can become the interface that can transform based on the user’s usecases - video editing, 3D modeling, Game develpment etc. etc. through native extensions or MCP, it is totally possible to name a few…
And yes I’ve used Claude Opus 4.6 with antigravity for Unity Development and no google or other model was able to perform the tasks as opus used to… I haven’t tested gemini 3.5 flash yet for same purpose.
I am also waiting for Gemini 3.5 Pro - if it would be able to compete with Fable 5 or if google announces their own version of this kind…
Antigravity offers the Claude model, which has two benefits: one is that multiple models help make the IDE bigger and stronger, and the other is that users’ projects built with Gemini can be improved through the better Claude model, providing the model team with better distillation data.
don’t think so because simply its just too expensive to add as a plan based because it is already too expensive per million tokens but they have it in the google cloud so you can just use google vertex from google cloud to just add fable 5 if you want it so much but in my opinion fable 5 is not such a big game changer because anthropic doesn’t provide us the non bussiness users with like complete fable 5 that is the same on the benchmarks but more limited one and in my opinion that model is the same as claude opus 4.6 but little bit “less” in some way I can say but you still can achive claude fable 5 level performance with simple prompting and skills
Waiting for Mythos/Fable 5 since it looks like the next major Claude release. Anyone heard anything?