codex - 💡(How to fix) Fix Allow compaction to use a different model [1 comments, 2 participants]

TL;DR

Configure Codex to use a cheaper model like gpt-5.4-mini when context exceeds 10% to improve efficiency.

Guidance

• Investigate the Codex configuration options to see if there's a built-in feature to switch models based on context size.
• Check the documentation for any existing compacting or model-switching mechanisms that can be leveraged.
• Consider reaching out to the Codex development team or community to ask about potential workarounds or future feature plans.

Notes

The feasibility of this feature request depends on the current capabilities and limitations of the Codex CLI.

Recommendation

Apply workaround: Since it's unclear if this feature is currently supported, exploring alternative solutions or workarounds, such as scripting or external tools, might be necessary to achieve the desired functionality.