Repository navigation
[BUG] GLM 5.1 stops reasoning and only provides direct output after /compact command #28623
Closed as not planned
Description
Description
After executing the /compact command to summarize the conversation context, the GLM 5.1 model stops performing its reasoning process. The model bypasses the "thinking" stage and immediately generates the final response. This behavior occurs consistently after the context has been compressed.
Plugins
No response
OpenCode version
1.15.6
Steps to reproduce
1.Start a session using the GLM 5.1 model.
2.Provide a complex prompt that triggers the model's reasoning/thinking phase.
3.Run the /compact command to compress the session history.
4.Submit a new follow-up prompt.
5.Observe that the model provides the output directly without any reasoning process.
Screenshot and/or share link
No response
Operating System
WSL Ubuntu
Terminal
No response
Metadata
Metadata
Assignees
Labels
No labels
Activity
github-actions commentedon May 21, 2026
This issue might be a duplicate of existing issues. Please check:
If your issue is distinct (e.g. reasoning is silently skipped rather than erroring or rendering incorrectly), feel free to clarify and this can remain open.
Arcadi4 commentedon May 21, 2026
Intercepting your request with this tool might be helpful for debugging. There might be some fields in the request being discarded. Please be aware not to upload any sensitive information.
github-actions commentedon Jul 21, 2026
To stay organized issues are automatically closed after 60 days of no activity. If the issue is still relevant please open a new one.