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
{{ message }}
Repository navigation
opencode run command uses incorrect model after /models selection #736
I used the /models command in the TUI to select a specific model.
When re-entering the opencode TUI, the program correctly uses the previously selected model.
However, when I directly execute the opencode run command (outside the TUI), it uses a different model than the one I selected.
I will attach screenshots below showing the model used in the TUI and the model used when running opencode run directly.
The issue is a feature actually haha. The thing is that when you select a model in the TUI it is stored in a different file to be used later as the recent used model. But that only works for the TUI.
Activity
Yusuf007R commentedon Jul 7, 2025
@xyb do you have any default model in the config.json?
xyb commentedon Jul 8, 2025
@Yusuf007R No. The config is attached below:
cravenceiling commentedon Jul 8, 2025
The issue is a feature actually haha. The thing is that when you select a model in the TUI it is stored in a different file to be used later as the recent used model. But that only works for the TUI.
xyb commentedon Jul 9, 2025
@cravenceiling This behavior is unexpected. We should add a note to the run command's documentation to clarify it.
xyb commentedon Jul 9, 2025
And the Gemini CLI displays the default model name, which is more user-friendly. We can consider it as a reference. See below:
rekram1-node commentedon Sep 5, 2025
will be fixed in next release
Merge pull request anomalyco#736 from pvgomes/docs/add-brazilian-port…
feat(app): virtualize session message rendering (anomalyco#736)
feat(apps): iris apps create "<need>" --node — run a Hive App kit you…