that’s 120% true, the speed of consuming tokens is crazy, its something like a trap?
I just simply unsubscribed Google AI Ultra. I can just use up the Claude Opus 4.6 session quota by simply about 10 turns in Antigravity. Before this, Antigravity is the best IDE that I can polish something really need human craftsmanship but also requires AI agent assist in process mass inputs. But now, I just went back to VS Code with the chat panel from Claude Code.
This should make an interesting poll.
why you have time to edit my comment but dont have time to make a public announcement? thats all we need after all?
It’s time to leave. They won’t care even if the subscription is still active.
Ultra too, serious problems too. But the waiting time is 5 hours (not 4).
How is your experience compared to antigravity?
Also, how much Opus 4.6 usage do you get with Claude Code compared to Antigravity?
idk but I am in Ultra on Antigravity and one task with sonnet consumed 60% of my quota. it cannot get worse than this.,
I was considering moving to Claude Code, but they also seem to have quota issues. And i think current issues we are facing in Antigravity Ultra plan is related to that exact problem.
https://www.reddit.com/r/ClaudeAI/comments/1s6yv86/20x_max_usage_gone_in_19_minutes/
Can anyone confirm this? If Claude code fixed that “?bug” I will move to claude code right away.
I have ultra business in work, and ultra personal at home - spent the day working with Opus without even making a dent in the usage. Finish work and make 1 prompt on ultra personal and instantly used up, from untouched. Also says refresh in 1 day and 5 hours. Bought ultra personal 10 days ago… Feels like the RAM situation again, give businesses maximum products, give single users paying the same amount nothing
It’s not only a Claude problem
Gemini Pro 3.1 quotas also downgraded dramatically on Ultra plan also. I just burned it in 1h and 20 minutes during trivial task. At the same time they changed ToS and silently introduce weekly limits for all plans including Ultra
so looks like it’s not a bug, it’s a new reality
oh no. weekly limits on ultra will kill them
could you test if max_thinking_length is set to 1024 tokens on business plan as well?
I am getting exactly the same issue. Felt a 5 x reduction of my quota from last week to now. Its a ridiculous change to be implemented without proper explanation.
My Gemini Pro 3.1 is working as before, hit limit after using 4 hours of heavy usage. Opus is burning in 1 hour
I love burning thru all the tokens in a hour due to multiple request failures for 1 prompt on ultra on top of constant hanging terminals. Can’t wait to see what they ship next to make my workflow slower.
Thanks for this, maybe I will switch my ultra account over to my business account. I been using it on my personal account all this time. Unfortunately my personal just renewed
I am on paid Pro.
4 prompts with Gemini 3.1 Pro (Low) in Fast mode and I am done for 6.5 days. Claude went even faster. I am not even trying Planning mode. This is a total joke. I’m already looking for a different solution and drop this nonsense.
I feel like we are being driven out of this product or lower-paid tiers on purpose. Google remains quiet about it; they are just watching us twitch like ants.
Sir, any good news or official anouncement?
7 hours ago, the refresh interval for gemini pro was 3 hours 33 minutes and claude models was 1 hour 51 minutes; and both have around 60% quota left before sleeping. When I woke up, and open my mac, now seeing 3 days 13 hour for gemini and 6 days 13 hours for claude without giving a single character prompt. @Googol_Admin @Abhijit_Pramanik
