Model not available error

Codex Codex CLI
,
1

When a model hits capacity in the middle of an active task, the interruption is very disruptive. The system should ideally preserve continuity automatically—either by letting the current task finish, transparently switching to an available compatible model, or handing off the full working context without requiring the user to restart or reconstruct the task.

Capacity errors are understandable, but they should be handled differently once a task is already in progress.

2

Hi @suman2, I understand that hitting a model capacity limit in the middle of an active task can be disruptive, and that there should be a clearer way to preserve task continuity when this happens.

I’ll pass this along to the team as feedback for consideration.

Thanks for sharing your experience!

~ Smith