Repository navigation
stream disconnected before completion: Transport error: network error: error decoding response body #8302
Description
What version of Codex is running?
0.75
What subscription do you have?
pro
Which model were you using?
gpt-5.2-codex-xhigh
What platform is your computer?
No response
What issue are you seeing?
It seems the service was outage now.
What steps can reproduce the bug?
It seems the service is outage now. I just say hi, and model not response.
What is the expected behavior?
No response
Additional information
No response
Activity
github-actions commentedon Dec 19, 2025
Potential duplicates detected. Please review them and close your issue if it is a duplicate.
Powered by Codex Action
etraut-openai commentedon Dec 19, 2025
We're not aware of any service outages currently. Please use the
/feedbackcommand to upload your logs and session details, then post the thread ID here.Are you using a VPN or any network proxies?
SunRunAway commentedon Dec 20, 2025
met same problem today：Uploaded thread: 019b3921-f0df-7fa1-a471-866e632e6461 gpt-5.2-codex-high
Ivan-composer commentedon Dec 20, 2025
Same issues for 5.2-codex in v0.75.0. MacOS 15.6.1
5.1-codex-mini works just fine, 5.2 high too.
019b3782-6b2b-7983-9153-18f4f91bddbb
neoAimeow commentedon Dec 20, 2025
same as 019b3842-bf50-7970-80e9-e979eb43291a
neoAimeow commentedon Dec 20, 2025
When running multiple Codex CLI processes at the same time, only one process is able to execute, and it performs very slowly. The others all fail with errors：
stream disconnected before completion:Transport error: network error: error decoding response bodyHowever, as a Max plan user, I inevitably need to work on multiple projects concurrently.
SunRunAway commentedon Dec 20, 2025
Starting from yesterday
niksteel123 commentedon Dec 20, 2025
same issue
jpcweb commentedon Dec 21, 2025
Same here with Codex 0.77.0 and gpt-5.2-codex.
TrueBers commentedon Dec 21, 2025
Same today. v0.76.0 on MacOs, gpt-5.2-codex high. Completely unusable.
44 remaining items
etraut-openai commentedon Jan 13, 2026
We think we have mitigated this issue. There are times when the "beefier models" (to quote @fmhall) take longer to generate the next batch of tokens during streaming, and this can sometimes exceed an HTTP timeout value.
I'm going to mark this bug as closed, but if you see this again in the next few days, please use
/feedbackto upload your logs and post your thread ID here.Radar-Lei commentedon Jan 16, 2026
The problem still exists.
ID: 019bc51e-a262-77d2-914e-64ea64054e0f
Kritoooo commentedon Jan 16, 2026
Same problem.
etraut-openai commentedon Jan 16, 2026
@Kritoooo, please use
/feedbackto upload your logs and post the thread ID here.Kritoooo commentedon Jan 18, 2026
019bcedb-5103-78c1-95ec-b5709efa3f72
Trojanking123 commentedon Jan 21, 2026
019bdf3b-39d9-7a50-bcf3-db715a50a8cf
etraut-openai commentedon Jan 21, 2026
@Kritoooo, your logs don't show a "Transport error". They show that the agent attempted to access the network but was blocked due to sandbox policy.
@Trojanking123, thanks for attempting to upload your logs. Unfortunately, they didn't come through. We had a temporary server issue that prevented some logs from uploading. We've resolved that issue now. If you're still able to repro this problem, please use the
/feedbackcommand again to upload your logs and create a new bug report. Sorry for the inconvenience.I'm going to close this bug report thread because we think we've mitigated the original bug that was reported. If you are seeing similar behaviors, it's likely a different issue, and we'd appreciate it if you'd file a new bug report with details.
MisterRound commentedon Jan 26, 2026
This issue has sadly returned. I've had a rough 48 hours over this, I get a ~12-15 output token delay from first input and multiple stream transport errors. I've got a ton of session ID's with this error, most recent is 019bf7ad-9cdc-7e30-a9e1-517068e0a6c4, using v0.91
etraut-openai commentedon Jan 26, 2026
@MisterRound, the underlying cause of this bug report was mitigated. If you're seeing a similar problem, then it's a new issue. Please use
/feedbackto upload your logs and session details and open a new bug report.