latest copilot in vscode loses its mind after a while.

[ibeyer-evotek]

🏷️ Discussion Type

Bug

💬 Feature/Topic Area

VS Code

Body

VS Code Copilot Chat fails with HTTP 400 after session has been active for a while

Summary

After the latest update to the Copilot Chat agent in VS Code, Copilot Chat requests begin failing once a chat session has been active for a while.

The issue does not appear immediately in a fresh session. Instead, it seems to emerge only after the session has accumulated enough history/context over time. Initially, reloading or relaunching VS Code would temporarily work around the issue, but that no longer fixes it.

This does not appear to be tied to rate limits or token usage. The failure mode is a request validation error from the backend, not a limit-exceeded or context-length error.

Error

Sorry, your request failed. Please try again.

Copilot Request id: 72c2b87f-c5d6-4f58-ab33-a4b4bc9c548f

GH Request Id: E261:3FFC99:518E66C:59BE5C6:6A073096

Reason: Request Failed: 400 {"error":{"message":"Invalid 'input[101].id': 'thinking_0'. Expected an ID that begins with 'rs'.","code":"invalid_request_body"}}

Expected behavior

Copilot Chat requests should continue to succeed even after a session has been active for a long time and accumulated substantial conversation history.

Actual behavior

After a session has been going for a while, Copilot Chat begins failing consistently with HTTP 400 and becomes unusable in VS Code.

Notes

• This started after the latest Copilot Chat agent update.
• The issue appears only after a session has built up history/context over time.
• It does not seem to be tied to rate limits or token usage.
• Restarting/reloading VS Code used to temporarily fix it, but no longer does.
• The backend rejects the request because input[101].id is thinking_0, while it expects an ID beginning with rs.
• This suggests a regression in request serialization or conversation-history replay, where an internal "thinking" item is being included in the outgoing request with a non-canonical ID format.

Environment

• VS Code version:
• GitHub Copilot version:
• GitHub Copilot Chat version:
• OS:

Reproduction

1. Open VS Code.
2. Start using Copilot Chat normally.
3. Continue the same session long enough for substantial chat history/context to accumulate.
4. Submit another request in that same session.
5. Observe the HTTP 400 failure.

Additional context

The error appears consistent with a malformed request body generated from persisted or replayed session history, rather than with quota exhaustion, token limits, or normal transient service failure.

[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. ⭐

[RAAJK20Pro]

This strongly looks like a regression in how Copilot Chat handles persisted conversation state or session replay after long-running chat sessions.

The key detail is the backend validation error:

"Invalid 'input[10].id': 'thinking_0'"

That error is particularly interesting because it does not resemble a normal token-limit, quota, or rate-limit failure. Instead, it suggests the client eventually sends an invalid internal message identifier while reconstructing or replaying accumulated chat history.

A few things from your report point toward a serialization/state-management issue rather than normal context exhaustion:

• the problem only appears after substantial session history accumulates
• requests initially work normally
• restarting VS Code previously worked temporarily
• the backend explicitly rejects request structure/validation
• the failure becomes deterministic over time
• the error references an internal-looking identifier (thinking_0)

The thinking_0 identifier especially stands out because it looks more like an internal reasoning/tool state object than a normal persisted chat message ID. It is possible that:

• hidden/internal reasoning entries are leaking into serialized chat history
• session compaction/truncation logic is malformed
• replayed conversation state includes unsupported internal message objects
• the latest Copilot Chat update introduced a mismatch between client-side and server-side expected ID formats

Another possibility is that the agent/history persistence layer is accidentally retaining transient internal nodes that should never be included in outbound request payloads.

The fact that reloading VS Code no longer resolves it may indicate:

• corrupted persisted workspace chat state
• cached session replay surviving restarts
• extension storage retaining malformed conversation objects

Possible troubleshooting steps

• Start a completely fresh Copilot chat session
• Clear Copilot chat/workspace state if possible
• Test in a clean VS Code profile
• Compare Stable vs Insiders builds
• Temporarily disable other AI/chat-related extensions
• Test whether the issue reproduces faster in agent/tool-heavy conversations
• Check whether specific workspaces trigger it more consistently

Additional useful diagnostics

It may help to include:

• exact VS Code version
• Copilot extension version
• OS/platform details
• whether chat history persistence is enabled
• whether MCP/tools/agent features were used heavily
• whether the issue occurs across multiple repositories/workspaces

If developer logs are available, capturing the outgoing payload shape around the failing request could probably confirm whether internal thinking_* nodes are being serialized incorrectly.

Overall, this looks much more like a client-side conversation-history serialization regression than a standard HTTP 400 caused by user prompts, token limits, or quota exhaustion.

The reproduction details and backend error context you provided are actually very useful for debugging 👍

[VictorMinardAtSiemensEnergy]

I can second this, I have been experimenting the exact same issues, though the message linked to the 400 return sometimes varies. Right now, I'm getting:
Reason: Request Failed: 400 {"message":"messages.17.content.11: Invalid data in redacted_thinking block"}

VS Code 1.120.0
github.copilot-chat Version 0.48.1
OS: RH UBI 9.4 (devcontainer inside WSL RedHat 10.1 inside W11 25H2)

[yutotakano]

I had the same issue but resolved it by explicitly choosing an LLM model in the chat window (e.g. "Raptor mini"), instead of keeping it as "Auto". It appears to be related to microsoft/vscode#316536, for which a fix is released to VS Code Insiders.