It seems that after updating to Claude Opus 4.6, the quota has become very wasteful. Even though I have only used it a few times, it has already reached its limit. I also experience a long processing time when the agent processes my commands.
1 Like
Same issue faced. My quota limit exceeded by giving just 1 prompt. I was like seriously, Opus 4.5 was better than this. At least i was able to use it for 3-4 hours. The google should either fix this issue or allow the users to use Opus 4.5.
4 Likes
Same problem here. I’m using Google Pro plan.
yes there are meticulllasly making sure the ai makes more mess quota exhauste even i 1 hr and now cluade model is on time out for 5 days strday only half day i worked dint even touch the claude today mrng im seeing 5 days lockout too much meess been made these days.. pro plan is more disatrusrous… ai going rouge deleteing the working block.. tokens gone for fix we have wait tiill next refresh…. such a clever marketing statergy to pull more subscritptions all month… but branding as antigravity is the best in 2026.. google made huge record breaking revenue but wats the use still ur product is such a mess !!
Hello,
Thank you for sharing your detailed feedback regarding the change in quota consumption after updating to Claude Opus 4.6.
The increased token usage observed with Opus 4.6 is primarily due to the model’s advanced reasoning architecture. There are a few key factors contributing to this:
- Claude Opus 4.6 is a frontier reasoning model that utilizes an internal process to solve complex problems.
- Opus 4.6 has a significantly larger and more capable context window.
- As mentioned in our documentation, the platform maintains “stickiness” for reasoning models to ensure consistency across a chat session. This means the high-capacity model stays active for follow-up questions.
Yes, i agree with that but think from our perspective also, not able to use the model for 4-5 days really makes a mess. Our most of the development process gets stopped. I am not saying you are wrong but there should be a middle solution for this problem. Because we shifted to anigravity just for the AI models. And now if they are not working properly then there is no meaning of shift from other IDE’s to Antigravity. Just think about this topic and i request to team to resolve this debate.
There is a much larger thread on this very topic - and this doesn’t really address the 5-10+ day lockout. I’d rather have a 5 hr lockout every day or even twice a day than run for a few days and then be locked out for 5+ days - this is really bad UX. it’s rough for development flow.
1 Like