Hi everyone,
Is anyone else getting completely locked out of their workspace for 11+ hours under the new Antigravity 2.0 update?
Ever since the shift to the shared Compute-Used system, using Gemini 3.5 Flash—which used to reliably refresh every 4 to 5 hours—now triggers massive, arbitrary cooldown periods. A single complex task or background agentic loop completely drains the weekly baseline cap, and the system immediately demands enabling “AI Credit Overages” to continue.
Waiting 11 hours or more for a baseline quota refresh makes it impossible to rely on this platform for daily production and analysis.
The product team needs to hear us on this: we need separate quotas for the Flash models and a hard limit on how much background agentic tasks can eat into our user-facing quota.
Please upvote this if you are experiencing the same ridiculous wait times!
Make sure to attach the screenshot from your prompt to both the email and the forum post, as visual proof always pushes support tickets higher up in priority!