Execution failed: request ended without sending any chunks

[williamromero]

Describe the bug

Everytime I ask for a premium reuqest, I'm obtaining this error:

   ↪ 2 lines...
 ✗ Execution failed: request ended without sending any chunks

   ↪ 199 lines...
 ✗ Execution failed: request ended without sending any chunks

This is interrupting every single request I do stoping the results I'm asking and I'm waiting for.

Affected version

0.0.351 Commit: 90770bc

Steps to reproduce the behavior

No response

Expected behavior

Obtain the completed response I'm asking to the model.

Additional context

Operating System: MacOS
CPU Technology: ARM
Terminal emulator: VS Code / zsh

[AungMyoKyaw]

I am also getting this kind of error. it is wasting my premium request count. 😢

[EnriqueGF]

Same:

Execution failed: request ended without sending any chunks

And I'm paying for this.

[Tachenpie]

Same. It looks like I have to quit it and restart and lose all my context.

[m5g5qxvxgb-maker]

did smbd solved it?

[EvanBoyle]

Does anyone have a consistent repro for this, or is this an intermittent issue? And is this haiku specific, or are you seeing this across models?

[gpltaylor]

I had the same issue today.

0.0.358
Commit: f5a8b1e76
Terminal: Powershell 

I copied and pasted some data into the prompt and this caused an error. When restarting following closing the terminal it worked. When I restarted wtihin the same active terminal it kept failing.

[tuminoid]

Does anyone have a consistent repro for this, or is this an intermittent issue? And is this haiku specific, or are you seeing this across models?

I'm getting this now with Opus 4.5. My current session has plenty of (cached) input, but not many requests consumed.

 ✗ Execution failed: request ended without sending any chunks

 ● Total usage est:       15 Premium requests
   Total duration (API):  21m 38.386s
   Total duration (wall): 1h 37m 12.426s
   Total code changes:    0 lines added, 0 lines removed
   Usage by model:
       claude-opus-4.5      12.6m input, 47.8k output, 11.3m cache read (Est. 15 Premium requests)

Suggesting it to continue results just in same chunk error.

 ➜ copilot --version
0.0.369
Commit: 83653a1

Ubuntu 24.04.03 LTS on x86, if it matters.

[examon]

This was taken care of in v0.0.407. Streaming responses now automatically retry when interrupted by server errors, so transient failures no longer immediately surface as the "Execution failed" error. If you're still seeing this consistently on a recent version, please feel free to reopen.