This is getting ridiculous

[TMRomain]

Select Topic Area

Product Feedback

Copilot Feature Area

Copilot Agent Mode

Body

I'm on the Copilot Pro plan, which promises "unlimited" access to GPT-4.1 as the base model. Yet, after only four agent-chat prompts, I receive the message: "You’ve reached the rate limit. Please wait or switch models." and it dosn't come back even after like 10 minutes of waiting or if it come backs I still get rate limits after 1 or 2 request.

Copilot has gone from being one of the best and most affordable services to one of the worst in just a few months. I was praising it only three months ago, but seeing what the competition now offers, I have to wonder why anyone would bother with Copilot. It’s disappointing to have to switch to a third-party editor when I prefer the simplicity of VS Code.

I understand that the terms mention "high demand may result in rate limiting," but what am I paying for if I can't even work uninterrupted for a few hours due to these limits ? Am I paying for random service availibility ?

The futur introduction of the premium requests also seems like a downgrade, although I do hope it addresses the current rate-limiting issues.

[Akash1134]

Hello @TMRomain!

Thank you for sharing your feedback about GitHub Copilot Pro's rate limiting with Agent Mode. I understand your frustration with hitting rate limits after only a few agent-chat prompts, especially as a paying Pro subscriber.

Regarding rate limits for Copilot Agent Mode:

According to GitHub's official documentation, while Copilot Pro provides unlimited code suggestions, certain advanced features like Agent Mode are still subject to rate limits during high-demand periods. The system is designed to ensure fair usage across all users.

What you can try:

1. Switch models temporarily: When you encounter rate limits, try switching to a different base model as suggested in the UI message. According to the VS Code Copilot documentation, this can help you continue working while waiting for rate limits to reset.

2. Optimize your prompts: Breaking down complex tasks into smaller segments can help reduce model usage and minimize rate limit encounters.

3. Check your connection: Sometimes intermittent connectivity issues can trigger false rate limit messages. Ensuring a stable connection may help.

I'm sharing your feedback with our product team, as your experience highlights important concerns about the expectations set for Pro subscribers. The upcoming premium requests feature is being developed to address these kinds of usage patterns.

For immediate assistance with persistent rate limiting, please consider contacting GitHub Support with details about your specific usage patterns to help them investigate.

Thank you for your patience and for being a Copilot Pro subscriber.

Best regards,

[Beechside]

I, too, am a paying Copilot Pro subscriber. I have to swap agents all the time and so chat is all but unusable. Typically just a couple of requests and the limit hits. I want an integrated code/suggest/chat dev environment.

If nothing improves, I'm going to cancel my subscription and go elsewhere...

[vanneszias]

Same over here. Getting rate-limited NON-STOP. I cannot do a coding session without getting limited each hour. It's getting ridiculous. If anybody here has some info about what to do to make this more bearable, PLEASE let me know.

[TMRomain]

The last few days the felt great for me, I could use Claude 4 sonnet almost all day without hitting the rate limit, it should always be like that.

[chrischne]

I am also a paying Copilot Pro subscriber and I am getting rate limited even on GPT-4.1, which supposed to be the base model as far as I know. This is ridiculous GitHub.

[David-Ben-Mesecke]

If I were to upgrade to pro+ or enterprise, would that remove that rate limit, cause its breaking the code I build with the tool making it useless.