HTTP 429 RESOURCE_EXHAUSTED on all models despite 100% quota shown (Starter Quota, Google AI Plus)

Every agent request fails immediately with “Agent terminated due to error”: HTTP 429, RESOURCE_EXHAUSTED, “Resource has been exhausted (e.g. check quota)”. Prompt: “reply with ok”, new conversation each time. Signing out and in did not help.

  • Models tried: Claude Sonnet 4.6 (Thinking), Claude Opus 4.6 (Thinking), GPT-OSS 120B (Medium). All fail the same way.
  • Settings > Models shows 100% remaining for all three. AI credits: 0.
  • Plan on Account page: Antigravity Starter Quota (Google AI Plus).
  • OS: Ubuntu Linux x64. antigravity --version: 1.107.0, commit 15487b3041e65228cae24980a3f796c905ef582c. An update (2.0.6) is available but not installed.
  • Trajectory ID: fde4ff84-7e43-4348-b044-ae602006127a / TraceID: 0x2aaaed092bda54..
  • Trajectory ID: 60786409-d5c1-4726-a1b6-325e440275b1 / TraceID: 0xf647f4df209efd..

Please check whether an account entitlement or quota-sync issue is blocking requests for this account.

Hi @Kinshuk_Gupta1,

Why This Is Happening Google AI Plus includes the base Antigravity Starter Quota ( Plans | Google Antigravity Docs ). Because you’re on an older build (1.107.0), your local Settings > Models page is showing a stale 100% remaining bar for Claude 4.6 and GPT-OSS 120B, even though the server is blocking them with HTTP 429 RESOURCE_EXHAUSTED because you have 0 AI credits / Starter Quota limits on those models ( Choose a model | Google Antigravity Docs ).

Since a normal sign-out/sign-in didn’t clear it, try these quick steps:

If it still fails after updating and resetting onboarding, please send over your diagnostic logs in a .txt file (from Help > Toggle Developer Tools > Console or View > Output > Antigravity Language Server) or a screenshot!