handle (Gemini) rate limits more gracefully

I selected Gemini 2.5 Flash (which I believe is a free model), and am regularly seeing this:

However after waiting a short while, I can submit a new prompt and continue without any issues. This makes sense because according to https://ai.google.dev/gemini-api/docs/rate-limits, the free model is limited to 10 requests per minute.

So ideally opencode would:

know the rate limits for each model
keep track of how close its usage is to that limit
warn when it's getting close to the limit
when the limit gets hit, pause for the appropriate amount of time showing a countdown timer and an explanation of what's going on together with some suggestions on how to overcome the limit (e.g. by changing to a paid model)

I'm actually getting the same for Claude, should I open new issue?

EDIT: it seems that this only applies for using anthropic through API key not oauth?

I also have this issue with openrouter

[automated] Closing due to 90+ days of inactivity. Feel free to reopen if you still need this!

Navigation Menu