Repository navigation
After changing model_providers, the resume command cannot view the conversation history of the previous provider #9054
Description
What version of Codex is running?
v0.80.0
What subscription do you have?
After modifying model_providers, the resume command cannot view the conversation history of the previous provider
Which model were you using?
No response
What platform is your computer?
Linux 5.15.0-60-generic x86_64 x86_64
What issue are you seeing?
I am using the APIs of different model providers to use Codex. After switching to a different model provider, I want to resume the previous conversation, but I find that the output of the resume command does not contain my previous records. I have to switch back to the previous model provider to restore my conversation history.
Note: I used resume to view the conversation history in the same directory.
What steps can reproduce the bug?
Switch the model provider and then use the resume command to view previous chat history
What is the expected behavior?
I hope that all conversation records can be shared between different model providers rather than being bound to a single model provider
Additional information
No response
Activity
[-]After modifying model_providers, the resume command cannot view the conversation history of the previous provider[/-][+]After changing model_providers, the resume command cannot view the conversation history of the previous provider[/+]github-actions commentedon Jan 11, 2026
Potential duplicates detected. Please review them and close your issue if it is a duplicate.
Powered by Codex Action
NormanMises commentedon Jan 11, 2026
No, it's not the same.
etraut-openai commentedon Jan 11, 2026
Conversations cannot generally be shared between model providers. This is by design. There are too many differences between models and their assumptions.
NormanMises commentedon Jan 12, 2026
But are there options to export conversations and import them?
etraut-openai commentedon Jan 12, 2026
No, that's not really practical. Conversations include details like CoT (chain of thought) tokens that do not transfer between models. You should stick to one model per thread. If you want to use a different model, create a new thread.
NormanMises commentedon Jan 13, 2026
Alright, thanks. ❤️