Error messages

You've reached your weekly rate limit.

You've reached your weekly rate limit.

GitHub Copilot CLI stops the session with this when your account has made as many model requests as GitHub allows in a week. A report from April, on the copilot-cli repo, shows the whole line:

You've reached your weekly rate limit. Please wait for your limit to reset on April 20, 2026 at 2:00 AM or switch to auto model to continue. Learn More (https://docs.github.com/en/copilot/concepts/rate-limits).

We read the pieces it's built from in Copilot CLI 1.0.88, which keeps its code compressed inside a single binary. The reporter wanted to know whether every model was blocked and whether they could carry on from where it stopped.

Four rate limits, four openings

The CLI has a different first sentence for each kind of limit, and the one you see tells you how wide the block is:

• You've hit the rate limit for this model. Only the current model is capped, and the advice is "switch models or wait for your limit to reset".
• You've hit your session rate limit. Before 1.0.34 this one said "global rate limit", and the changelog notes the rename.
• You've reached your weekly rate limit. The long one, with a reset date that can be days away.
• You've hit your rate limit. The generic fallback.

The weekly line ends with or switch to auto model to continue, so not every model is blocked: auto (Copilot's mode that picks a model for you) is still open. The session itself survives, so you can carry on in it once you've switched.

What to do

When the limit hits, the CLI asks "Switch to auto mode?" and offers Yes, "Yes (always)" and No. Yes moves this session to auto and keeps going. If you'd rather never be asked, the continueOnAutoMode setting (off by default, listed under copilot help config) makes the switch automatic for per-model, weekly and integration limits. Its own help text says it doesn't apply to global rate limits, generic 429s (the HTTP code for too many requests) or models you've connected with your own key.

If auto isn't good enough for the work, as the April reporter felt about planning tasks, the only other road is waiting for the printed time. GitHub's limits page suggests checking whether anything automated is sending rapid requests, and upgrading if you're on an individual plan.

When it doesn't lift on time

A Copilot Pro+ user in April, working in Copilot Chat, was still blocked more than an hour after the reset should've happened. Their error came from the server in different words: Sorry, you've exceeded your weekly rate limit. Please review our Terms of Service. with the code user_weekly_rate_limited. The issue is closed. If it happens to you, copy the request ID printed at the end of the CLI's message into your report.

Running out of credits is different

Rate limits sit beside a separate allowance of AI credits that each plan includes, and the CLI words that one differently: You've run out of your included AI credits for the month., with a link to manage your budget at github.com/settings/copilot/features. Switching to auto won't fix that. According to GitHub's docs, the ways on are upgrading the plan or setting a budget for additional usage.