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
Opencode is super slow on my computer. There is a file called snapshot which turns huge like 16 GB so I have to remove it regularly. Not sure if that is related. The built in models are extremely slow, replying every five minutes or so. But even paid models like those provided by nanogpt are equally slow. Everything has grinded down to a halt and the product is no longer useful. Please assist.
"Labyrinth, where she found the mucus of the green dragon. Vibrant colors. Relentlessly cool."
Reacted by Łukasz Byjoś, Asad Moosvi, Vo Duc Huy and adityagusain273
If this feature is not performant it reflects badly on opencode - users will think that either their model or generally the opencode system is slow. This should be off by default until performance improves.
Activity
github-actions commentedon Mar 7, 2026
This issue might be a duplicate of existing issues. Please check:
These issues cover the snapshot growing very large, which seems related to what you're experiencing.
Hona commentedon Mar 7, 2026
For now you can set snapshot: false in your config! :)
621625 commentedon Mar 7, 2026
In the opencode.json or elsewhere? What is the format and syntax?
"She travelled across the universe. Fast as a blix. Ready for adventures"
621625 commentedon Mar 7, 2026
Testing...
621625 commentedon Mar 7, 2026
Seems to work now. If not I will update again.
kozak-codes commentedon Jun 10, 2026
+1 Snapshots bad.
If this feature is not performant it reflects badly on opencode - users will think that either their model or generally the opencode system is slow. This should be off by default until performance improves.
rekram1-node commentedon Jun 23, 2026
Automated: closing question issues.
GitHub Issues are for bug reports and feature requests only — questions / support are no longer tracked here.
If this is a bug or feature request, please open a new issue with the appropriate template.