codex - 💡(How to fix) Fix When starting a new chat, thinking effort defaults to low and cannot be changed intil chat starts [1 comments, 2 participants]

TL;DR

The issue might be related to the configuration settings, specifically the model_reasoning_effort parameter in the config.toml file, which is set to "medium" but not reflected in the chat settings.

Guidance

• Review the config.toml file to ensure that the model_reasoning_effort parameter is correctly set and synchronized with the chat settings.
• Try changing the model_reasoning_effort parameter to a different value (e.g., "high" or "low") and see if the thinking effort setting in the chat updates accordingly.
• Verify that the model parameter is compatible with the desired thinking effort settings, as some models may have limitations or restrictions.
• Check if there are any other configuration settings that might be overriding the thinking effort setting, such as approval_policy or sandbox_mode.

Notes

The issue might be specific to the "gpt-5.5" model or the "Pro" subscription, and further investigation may be needed to determine the root cause.

Recommendation

Apply workaround: Try changing the model_reasoning_effort parameter in the config.toml file to a different value and see if the issue is resolved, as this is a simple and non-invasive change that can help identify the problem.