429错误

[kinomotom]

只能调用一次 第二次就429

[Kitoryu]

same problem. Cline 3.82.0

[wrangler:info] POST /v1/chat/completions 200 OK (7ms)
[GenerationConfig] Real thinking enabled for 'gemini-2.5-flash-lite' with budget: -1
Got 429 for gemini-2.5-flash-lite, retrying in 500ms (attempt 1/4)
Got 429 for gemini-2.5-flash-lite, retrying in 1000ms (attempt 2/4)
Got 429 for gemini-2.5-flash-lite, retrying in 2000ms (attempt 3/4)
Got 429 for gemini-2.5-flash-lite, retrying in 3000ms (attempt 4/4)

[ERROR] [GeminiAPI] Stream request failed: 429

"error": {
  "code": 429,
  "message": "You have exhausted your capacity on this model. Your quota will reset after 20s.",
  "status": "RESOURCE_EXHAUSTED",
  "details": [
    {
      "@type": "type.googleapis.com/google.rpc.ErrorInfo",
      "reason": "RATE_LIMIT_EXCEEDED",
      "domain": "cloudcode-pa.googleapis.com",
      "metadata": {
        "model": "gemini-2.5-flash-lite",
        "uiMessage": "true"
      }
    }
  ]
}

[ERROR] Stream error: Stream request failed: 429

[RyanGarber]

Getting 429s as well and I have a Pro sub. Weird because in my own implementation of Gemini CLI (through Vercel AI SDK's community provider) I had no such issue. Sadly don't have time to investigate. Perhaps they ratelimit Cloudflare IPs?

[ikechan8370]

Getting 429s as well and I have a Pro sub. Weird because in my own implementation of Gemini CLI (through Vercel AI SDK's community provider) I had no such issue. Sadly don't have time to investigate. Perhaps they ratelimit Cloudflare IPs?

The same thing happened even running locally via npm run dev. So Cloudfalre is not the reason. That's wired because I can use gemini cli on my machine without 429 problems.