Remote Compact Task Blocking GPT 5.3-Spark from Working After Usage Limit Reached.

Open 💬 2 comments Opened Aug 4, 2026 by antoineeliux
💡 Likely answer: A maintainer (github-actions[bot], contributor) responded on this thread — see the highlighted reply below.

What version of Codex CLI is running?

codex-cli 0.146.0

What subscription do you have?

Max

Which model were you using?

5.6 Max

What platform is your computer?

Macbook Pro M4

What terminal emulator and version are you using (if applicable)?

mac terminal

Codex doctor report

What issue are you seeing?

When I run out of usage during an auto compact, it is not letting me change the model to get-5.3 spark even though I still have 100% usage for 5.3 spark. Well it does let me change the model but it just says it ran out of usage. It states "Error running remote compact task: You've hit your usage limit". I suspect because the compact task is still stuck on the 5.6 model.

What steps can reproduce the bug?

Run out of usage during a compact while on 5.6, then switch the model to 5.3 spark (which still has usage) and try to prompt 5.3 spark.

What is the expected behavior?

_No response_

Additional information

_No response_

View original on GitHub ↗

2 Comments

github-actions[bot] contributor · 23 days ago

Potential duplicates detected. Please review them and close your issue if it is a duplicate.

  • #36738

Powered by Codex Action

nos1609 · 23 days ago

@antoineeliux OpenAI support already explained it previously that Spark uses non-Spark limits and model for context compaction, even if its not really obvious from the docs