Critical issues with Copilot / Pro+ service

[kryre]

🏷️ Discussion Type

Bug

💬 Feature/Topic Area

Issues

Body

I wanted to share my experience with Copilot and the Pro+ service because at this point it’s becoming really hard to rely on it for actual work.

The biggest issue I ran into was with model selection. In one of my sessions, Opus 4.6 (x30 fast) kept being used even though I explicitly switched to other models. The UI showed the change correctly, but only within the current interaction. As soon as I sent another message in the same agent session, it immediately switched back to the x30 model again. So in practice, the change didn’t persist at all.

Because of that, around 10% of my request quota disappeared very quickly without me intending to use that model. The only way I managed to stop it was by going into settings and completely hiding that model. Only then did it actually respect my selection.

There are also serious problems with session persistence. More than once I closed a project, came back the next day, and all chats and agent sessions were gone. No recovery, nothing. Just lost work.

On top of that, there are constant request failures. The “413 Request Entity Too Large” error shows up frequently and none of the models seem able to handle it properly. The only option is to keep retrying, which leads to another issue — rate limits.

Very often I get rate limited for no clear reason. Today is a good example. I was working for less than an hour, mostly retrying failed requests because nothing else worked, and suddenly I got blocked. For the past 10 minutes I’ve been stuck with:

"Chat took too long to get ready. Please ensure you are signed in to GitHub and that the extension GitHub.copilot-chat is installed and enabled. Click restart to try again if this issue persists."

Earlier I also saw errors like:

• "No activated agent with id 'github.copilot.editsAgent'"
• "Language models loaded: false (rate limit exceeded)"

So basically retrying broken requests ends up locking you out completely.

Overall, there are still a lot of bugs in Copilot:

• model switching not working reliably (resets within the same session)
• quota being consumed incorrectly
• sessions disappearing
• random rate limits
• agents failing to initialize
• frequent request errors
• and the app itself sometimes crashes

Right now it feels unstable enough that it can completely block work, which is exactly what happened to me today.

Is anyone else running into similar issues? Or found any reliable workarounds?

Guidelines

• I have read the above statement and can confirm my post is relevant to the GitHub feature areas Issues and/or Projects.

[github-actions[bot]]

💬 Your Product Feedback Has Been Submitted 🎉

Thank you for taking the time to share your insights with us! Your feedback is invaluable as we build a better GitHub experience for all our users.

Here's what you can expect moving forward ⏩

• Your input will be carefully reviewed and cataloged by members of our product teams. 
  • Due to the high volume of submissions, we may not always be able to provide individual responses.
  • Rest assured, your feedback will help chart our course for product improvements.
• Other users may engage with your post, sharing their own perspectives or experiences. 
• GitHub staff may reach out for further clarification or insight. 
  • We may 'Answer' your discussion if there is a current solution, workaround, or roadmap/changelog post related to the feedback.

Where to look to see what's shipping 👀

• Read the Changelog for real-time updates on the latest GitHub features, enhancements, and calls for feedback.
• Explore our Product Roadmap, which details upcoming major releases and initiatives.

What you can do in the meantime 💻

• Upvote and comment on other user feedback Discussions that resonate with you.
• Add more information at any point! Useful details include: use cases, relevant labels, desired outcomes, and any accompanying screenshots.

As a member of the GitHub community, your participation is essential. While we can't promise that every suggestion will be implemented, we want to emphasize that your feedback is instrumental in guiding our decisions and priorities.

Thank you once again for your contribution to making GitHub even better! We're grateful for your ongoing support and collaboration in shaping the future of our platform. ⭐

[roshhellwett]

The workaround I found for the model issue (like you mentioned) is to just hide the problematic model entirely from settings. It's a pain but it actually works.

For the 413 error, that one usually happens when the conversation context gets too long. Splitting into a new chat session or reducing the amount of code/text you're sending in a single message tends to help. Not ideal, but it at least gets you unblocked.

For the rate limiting after retrying failed requests — that's a known frustration. Retrying basically counts against your quota even if the request never succeeds. If you hit that wall, the only real fix is to wait it out (usually 10–30 minutes) before trying again. Restarting VS Code and re-authenticating the GitHub extension sometimes clears the "Chat took too long to get ready" error faster.

For the session/chat disappearing issue, I'd recommend reporting that one directly to GitHub Support since losing saved work with no recovery option is a data loss bug, not just a UX issue. Include your account details and approximate timestamps — https://support.github.com/contact

Honestly a lot of these feel like stability issues on GitHub's end rather than anything you're doing wrong. Hopefully they get addressed soon because Pro+ should be more reliable than this.

[kryre]

Yeah, I ended up doing the same thing and hiding the Opus 4.6 fast model in settings, and that does seem to stop the issue.

But the problem is that before I realized what was actually happening, I was convinced I was working on a different model. The UI showed that the model was changed, so there was no reason to suspect anything was wrong.

Because of that, I ended up losing around 10% or even more of my request quota in a very short time, without intentionally using that model at all.

So while hiding the model is a workaround, it doesn’t really solve the underlying problem, especially since it’s very easy to lose a significant amount of quota before noticing what’s going on.

I also forgot to mention another issue related to chat sessions.

When I leave an agent chat and try to come back to it later, the session often gets stuck in an infinite loading state. It just keeps loading and I’m unable to open that conversation at all.

So even if the session wasn’t deleted, it effectively becomes inaccessible, which means ongoing work is blocked anyway.

Has anyone else experienced this or found a way to recover such stuck sessions?

[kryre]

When I open a session that goes into infinite loading, there’s no way to exit it normally. The UI just stays stuck and I can’t navigate away from that chat.

The only way out is to restart the window or the entire application, which makes the whole experience even more frustrating.