Repository navigation
TOO MUCH Error running remote compact task #24449
Description
What version of the Codex App are you using (From “About Codex” dialog)?
26.519.41501 (3044)
What subscription do you have?
pro 100USD
What platform is your computer?
Darwin 25.5.0 arm64 arm
What issue are you seeing?
Ever since a certain release of version 5.5, I've frequently encountered this issue whenever the context length becomes even moderately long. From my experience, the probability exceeds 50%, making Codex completely unusable now. In previous versions, I never experienced this problem, even with equally long contexts and under the same network conditions. This issue suddenly appeared in a recent update and has been occurring extremely frequently ever since. I encounter it heavily every day, and it's driving me to the brink of breakdown.
Codex ran out of room in the model's context window. Start a new thread or clear earlier history before retrying.
Error running remote compact task: stream disconnected before completion: error sending request for url (https://chatgpt.com/backend-api/codex/responses/compact)
What steps can reproduce the bug?
Feedback ID: 019e5f50-13ea-7a70-b8c6-37b28d5ccff7
What is the expected behavior?
No response
Additional information
No response
Activity
github-actions commentedon May 25, 2026
Potential duplicates detected. Please review them and close your issue if it is a duplicate.
Powered by Codex Action
jshaofa-ui commentedon May 25, 2026
yiheng8023 commentedon May 26, 2026
The same to me！
zhangyutao12 commentedon May 26, 2026
The same to me！
kuangyonglin commentedon May 26, 2026
The same to me！
pro 100USD
mickeyjoes commentedon May 28, 2026
Same here, since yesterday non-stop in all my codex sessions :/
Pro $100
CLI v0.134.0
Patrix999 commentedon Jun 4, 2026
same here, basically made codex unusable
etraut-openai commentedon Jun 29, 2026
Thanks for reporting this problem. Until recently, Codex used a separate endpoint and server-side logic for compaction. We recently switched to a more robust approach where the compaction logic is moved locally using the same endpoint as normal turns. This eliminates "remote compaction" errors and increases the stability and reliability of compaction operations. If you see further problems with compaction, please use
/feedbackand open a new bug report.