Troubleshoot Models and Providers
Use these checks when a model is missing, a provider cannot connect, or a request reaches a model-related limit.
A model is missing
- Confirm the active organization under Settings → Profile.
- Open Settings → Models and review the available defaults.
- Open Settings → Providers and confirm the required connection.
- Reload Reasoning after changing organization or provider configuration.
Organization access and direct provider authentication can produce different model lists.
Reasoning effort is unavailable
The reasoning control appears only when the selected model exposes supported variants. Select another model or use the current model's default behavior.
A provider cannot connect
Repeat the supported authentication or credential flow under Providers. Check the output channel for a redacted error and verify network or organization policy without posting credentials in Chat.
A request reaches a limit
Review the active account budget, the model's availability, and the provider's current rate window. Wait, choose another permitted model, or contact the organization administrator as appropriate.
Starting another session does not bypass the same organization or provider limit. See Usage, Costs, and Limits.