Request failed due to a transient API error. Retrying...

Describe the bug

Hitting multiple Request failed due to a transient API error. Retrying... that lead to Sorry, you've hit a rate limit that restricts the number of Copilot model requests you can make within a specific time period. Please try again in 1 minute. Please review our Terms of Service.

Affected version

GitHub Copilot CLI 1.0.6.

Steps to reproduce the behavior

No response

Expected behavior

No response

Additional context

OS : Ubuntu
I tried updating the system, close copilot, ...
Model : Haiku

Switching to gpt-5-mini resolves the issue for now.

Still hitting, but not to a blocking point:

● Request failed due to a transient API error. Retrying...

● Request failed due to a transient API error. Retrying...

I have the same problem

same

Getting same errors today consistently on v1.0.7
Also getting same errors in vscode

seems to be general issue, see current comments on https://github.com/orgs/community/discussions/180092

getting same error

Same for me and it slows down a lot the agent work

had this error once and in VScode it said sth like rate limitation ongoing. I think it was after the model was looping somehow. after a few hours it was gone.

It happens in v1.0.11 too, gpt-5-mini seems to work, but still slow. It'll probably take a while until infra catches up with demand.

Happening now to me

It's happening to me too AND it consumes requests, that's kinda bad.

also happens with opus 4.5 on pro subscription.
pay 3x credits per requests that result in ● Request failed due to a transient API error. Retrying... and then are idling for minutes...

its a pure rip-off... as expected from our lovely microsoft

✗ Sorry, you've hit a rate limit that restricts the number of Copilot model requests you can make within a specific time period. Please try again in 5 hours. Please review our Terms of Service
(https://docs.github.com/site-policy/github-terms/github-terms-of-service). (Request ID: XXXXXXXXXXXXXXXX)

need a tutorial on how to prevent me from commiting a hatecrime

I have the same problem

The same issue here with GPT-5.3-codex

Im also constantly getting these. Fleet mode, Opus 4.6 with a business subscription. It seems to just timeout before retrying. Its double annoying as it actually costs you not only time but actual premium request to try again after it gives up.

If there is a rate limit that fleet hits then CLI should be smarter on how fleet works to understand and coordinate this.

Same issue here when asking the agent (using Opus 4.6) to compare 2 markdown files that are about 80KB in size:

● Request failed due to a transient API error. Retrying...

✗ Execution failed: Error: Failed to get response from the AI model; retried 5 times (total retry wait time: 98.05322941541459
seconds)
Last error: CAPIError: 503 {"error":{"message":"HTTP/2 GOAWAY connection terminated","type":"connection_error"}}

For me, deleting the ~/.copilot/ dir and reinstalling fixed it

For me, deleting the ~/.copilot/ dir and reinstalling fixed it

That worked, thanks! It did delete my session history but I can probably find a way to restore it

@j-fdion

for me my copilot-instructions file was corrupted with what appeared to be a bad merge with thousands of lines. was blowing out the context window and causing the LLM to reject the call. Also explains the repeated "Compacting Conversation" messages.

Thanks to @j-fdion for the idea to check markdown files.

So, my issue was too many .md files, it was bloating the context and my message was not included in the prompt. I deleted all my useless .md files and it works now. The error was not helpful at all to find this. Not saying it's the problem every person here is having, but worthwhile checking into it.

I still get it sometime, that's quite annoying.

I encountered this problem all the time since yesterday, and I suspect there's a shadow limit disabling you to use more although quota still not exceeded. I'm using Pro plan and claude sonnet 4.6. Very disappointed about this.

Encountered this multiple times.

└─➜ copilot --version
GitHub Copilot CLI 1.0.42.

● Request failed due to a transient API error. Retrying...

For me the issue was due to a corrupt session. I restarted my mac while the session was active, and upon continuing the same session I started getting this error.

Fix: start a new session with the prompt There is a sssion with summary: "xxxx" on which prompts are failing with "Request failed due to a transient API error. Retrying...". Can you analyse what the issue with that session is? If possible, read the conversation history, restore it in current session and delete that session

After this I was able to continue my conversation on the new session

Getting the same error message, especially if a subagent/ skills is working and github starts compaction.
Also while doing web searches

I hit the same issue using the CLI - version 1.0.59. on Mac. I run Opus 4.8.

Thanks for the report. We've shipped fixes in this area since this was filed (most recently in Copilot CLI v1.0.82), and we believe this may now be resolved. Could you retest on the latest release (Copilot CLI v1.0.103) and let us know whether you still see it? If we don't hear back in a couple of weeks, we may close this issue — you can always reopen it.

_{Triaged as part of a Copilot SDK backlog review, with help from GitHub Copilot.}

Navigation Menu

Repository navigation

Request failed due to a transient API error. Retrying... #2101

Describe the bug

Affected version

Steps to reproduce the behavior

Expected behavior

Additional context

3 remaining items

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Request failed due to a transient API error. Retrying... #2101

Description

Describe the bug

Affected version

Steps to reproduce the behavior

Expected behavior

Additional context

Activity

AmauMaill commented on Mar 17, 2026

peter-mw commented on Mar 18, 2026

KlotzJesse commented on Mar 18, 2026

aro-ssystems commented on Mar 18, 2026

GitMonsters commented on Mar 24, 2026

danparisi commented on Mar 25, 2026

rs38 commented on Mar 25, 2026

mkontsek commented on Mar 26, 2026

doj-scraper commented on Mar 27, 2026

j-fdion commented on Mar 27, 2026

behpunktli commented on Mar 28, 2026

hxgdzyuyi commented on Mar 31, 2026

jakubjezek001 commented on Mar 31, 2026

3 remaining items

chrisaut commented on Apr 3, 2026

brentkendrick commented on Apr 3, 2026

sczesny commented on Apr 3, 2026

brentkendrick commented on Apr 3, 2026

squickone commented on Apr 4, 2026

danparisi commented on Apr 12, 2026

RX-105 commented on Apr 23, 2026

madhushesharam commented on May 6, 2026

danish-kamal-sophos commented on May 11, 2026

EngBuilds commented on Jun 2, 2026

shuaibird commented on Jun 5, 2026

patniko commented on Oct 26, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions