Ability to set reasoning effort for automations

Resolved 💬 5 comments Opened Feb 19, 2026 by iandanforth Closed May 16, 2026

💡 Likely answer: A maintainer (etraut-openai, contributor) responded on this thread — see the highlighted reply below.

What variant of Codex are you using?

App

What feature would you like to see?

I would like to be able to set the reasoning effort for automations. Today it defaults to medium. Some automations I want to set to low, others to extra high. As an example of an extra high automation I want to run is an ongoing research project that I would like an agent to follow up on on a daily basis. The performance of Codex App at 'Medium' has been underwhelming for this purpose.

Additional information

_No response_

View original on GitHub ↗

5 Comments

wbdb · 6 months ago

config.toml:

model = "gpt-5.3-codex"
model_reasoning_effort = "high"

or:

codex exec --skip-git-repo-check -m 'gpt-5.2-codex' -c 'model_reasoning_effort=high' 'Something todo'

https://developers.openai.com/codex/cli/reference/

iandanforth · 6 months ago

Thanks @wbdb for your suggestion. This feature request is specifically for the App. Perhaps I should specify that my request would be for this reasoning effort selector to appear in the Automations creation / editing views of the app.

tyrelh · 6 months ago

Yes please. I'm also interested in this feature in the app.

quang-flexos · 5 months ago

I would love to set different model / reasoning effort for each automation task please 🙏

etraut-openai contributor · 3 months ago

This feature was added a while ago.