You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When calling /compact, it seems like prompt caching isn't being used. Regular messages in the same session do get cache hits, so the caching setup is working - it just doesn't seem to apply to /compact.
Plugins
None
OpenCode version
1.1.34
Steps to reproduce
Connect OpenCode to a model served by llama.cpp. May or may not be reproducible with other providers.
Have a back and forth conversation and confirm prompt caching is being used either by looking at the server logs, or judging by time to first token.
Run /compact command and observe prompt caching not get used.
Activity
github-actions commentedon Jan 24, 2026
This issue might be a duplicate of existing issues. Please check:
Feel free to ignore if none of these address your specific case.
docs(github): add remaining issue analyses and implementation guides
fix(opencode): add prompt caching to session compaction - Issue anoma…
fix(opencode): add prompt caching to session compaction - Issue anoma…
fix(opencode): add prompt caching to session compaction - Issue anoma…
docs(github): add remaining issue analyses and implementation guides
TomLucidor commentedon Feb 24, 2026
Please keep this updated whenever possible for those who uses local models
github-actions commentedon Apr 26, 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.