GitHub Copilot Chat – Multiple bugs with third-party Claude model integration (VS Code)
- Dominant language
- TypeScript
- Stars
- 193k
- Forks
- 42.4k
- PR merge metrics
- PR metrics pending
Description
Type: Bug
Subject: GitHub Copilot Chat – Multiple bugs with third-party Claude model integration (VS Code)
Summary:
Over the course of a single day, I encountered a series of related bugs when using Claude models (via GitHub Copilot Chat / Copilot's third-party Claude agent integration) in VS Code. These appear to be distinct but related failures in Copilot's Claude integration layer, not issues with my configuration.
Environment:
- Editor: VS Code
- Extension: GitHub Copilot Chat
- Subscription: GitHub Copilot (no separate Anthropic/Claude subscription)
- Model(s) affected: Claude (Opus/Sonnet, exact version selected via Copilot's model picker)
Issues encountered (in order):
1. Chat stuck indefinitely on "Evaluating" / "Analyzing"
Requests would hang indefinitely at the "Evaluating" or "Analyzing" stage with no response and no error message. This occurred consistently across all requests, not just complex ones.
Related: https://github.com/microsoft/vscode/issues/312526
Related: https://github.com/microsoft/vscode/issues/311840
2. 400 error: "thinking.type.enabled" is not supported for this model
Every request to the Claude model failed immediately with:
"thinking.type.enabled" is not supported for this model. Use "thinking.type.adaptive" and "output_config.effort" to control thinking behavior.
This indicates Copilot's Claude integration is sending a deprecated API parameter format that the current model no longer accepts. This is a client-side compatibility bug, not something a user can fix via settings — there is no UI option in Copilot to control this parameter.
Related: https://github.com/microsoft/vscode/issues/312290
Related: https://github.com/anthropics/claude-code/issues/49219
Related: https://github.com/anthropics/claude-code/issues/51439
Related: https://github.com/anthropics/claude-code/issues/63546
3. "Error: Error during execution"
After working around issue #2, subsequent requests to Claude models failed with a generic "Error during execution" message, while the same prompts worked fine with GPT-based models.
Related: https://github.com/microsoft/vscode/issues/303965
Related: https://github.com/microsoft/vscode/issues/300226
4. Chat session history lost when using Claude (third-party) models
Sessions using Claude via Copilot are not reliably preserved in the Sessions panel — reopening a session that previously used Claude causes the agent to lose all prior context and reprocess/re-read the entire project from scratch, effectively discarding the conversation history.
Related: https://github.com/microsoft/vscode/issues/295745
Workaround applied:
Enabling the following user settings routed Claude requests through VS Code's newer "Agent Host" subsystem instead of Copilot's own proxy layer, which resolved issue #2 (the thinking.type.enabled error):
"chat.agentHost.enabled": true,
"chat.editor.claude.preferAgentHost": true,
"chat.agents.claude.preferAgentHost": true
However, this setting change itself caused all pre-existing chat sessions to be invalidated, forcing a full restart of context for ongoing work (related to issue #4 above).
Impact:
- Claude models were completely unusable via Copilot for most of a working day.
- Multiple hours of lost productivity due to repeated failures across different error types.
- No official acknowledgment or fix timeline found on any of the linked issues as of this writing.
Request:
Given the number and severity of these issues, could you please:
1. Confirm whether these are known/tracked issues on your end.
2. Provide a timeline or status update on fixes for the thinking.type.enabled parameter mismatch (issue #2), as this makes Claude models unusable in Copilot entirely.
3. Advise on preventing chat session/context loss when using Claude models (issue #4).
Happy to provide additional logs or reproduction steps if helpful.
VS Code version: Code 1.127.0 (4fe60c8b1cdac1c4c174f2fb180d0d758272d713, 2026-06-30T10:52:33+02:00)
OS version: Windows_NT x64 10.0.19045
Modes:
System Info
|Item|Value|
|---|---|
|CPUs|Intel(R) Core(TM) i7-10700K CPU @ 3.80GHz (16 x 3792)|
|GPU Status|2d_canvas: enabled
GPU0: VENDOR= 0x10de, DEVICE=0x21c4 [NVIDIA GeForce GTX 1660 SUPER], DRIVER_VENDOR=NVIDIA, DRIVER_VERSION=32.0.16.1062 *ACTIVE*
GPU1: VENDOR= 0x1414, DEVICE=0x008c [Microsoft Basic Render Driver], DRIVER_VERSION=10.0.19041.5794
Machine model name:
Machine model version:
direct_rendering_display_compositor: disabled_off_ok
gpu_compositing: enabled
multiple_raster_threads: enabled_on
opengl: enabled_on
rasterization: enabled
raw_draw: disabled_off_ok
skia_graphite: disabled_off
trees_in_viz: disabled_off
video_decode: enabled
video_encode: enabled
webgl: enabled
webgpu: enabled
webnn: disabled_off|
|Load (avg)|undefined|
|Memory (System)|31.87GB (18.34GB free)|
|Process Argv|--crash-reporter-id 20e8c1b6-ae75-4de5-8f91-dfe8571e84b8 --crash-reporter-id 20e8c1b6-ae75-4de5-8f91-dfe8571e84b8|
|Screen Reader|no|
|VM|0%|
Extensions (12)
Name|Identifier|Author|Version
---|---|---|---
Dart|dart-code.dart-code|Dart-Code|3.138.0
Flutter|dart-code.flutter|Dart-Code|3.138.0
Prettier - Code formatter|esbenp.prettier-vscode|esbenp|12.4.0
C# Dev Kit|ms-dotnettools.csdevkit|ms-dotnettools|3.20.199
C#|ms-dotnettools.csharp|ms-dotnettools|2.140.9
.NET Install Tool|ms-dotnettools.vscode-dotnet-runtime|ms-dotnettools|3.1.0
Python Debugger|ms-python.debugpy|ms-python|2026.6.0
Python|ms-python.python|ms-python|2026.4.0
Pylance|ms-python.vscode-pylance|ms-python|2026.2.1
Python Environments|ms-python.vscode-python-envs|ms-python|1.36.0
vscode-pdf|tomoki1207.pdf|tomoki1207|1.2.2
Unity|visualstudiotoolsforunity.vstuc|visualstudiotoolsforunity|1.2.2
A/B Experiments
```
vsliv368:30146709
binariesv615:30325510
nativeloc1:31344060
dwcopilot:31170013
dwoutputs:31242946
copilot_t_ci:31333650
e5gg6876:31282496
pythonrdcb7:31342333
6518g693:31463988
aj953862:31281341
6abeh943:31336334
envsactivate1:31538736
editstats-enabled:31346257
478i5457:31544283
cloudbuttont:31379625
3efgi100_wstrepl:31403338
ec5jj548:31422691
cp_cls_t_966_ss:31526232
4je02754:31466945
c3h7c220:31478652
ge8j1254_inline_auto_hint_haiku:31490510
38bie571_auto:31478678
cp_cls_t_1081:31454832
conptydll_true:31498968
ia-use-proxy-models-svc:31452481
e9c30283:31461165
test_control1:31471004
46204921:31447328
capi_a_098a3278:31544979
idci7584:31464702
e3e4d672:31494082
ei9d7968:31496641
nes-extended-on:31455476
chat:31457767
8hig5102:31480529
89g7j272:31518289
7e187181:31503455
i2gc6536:31499202
52612955:31516516
ghj88844:31499326
ddid_t:31478206
hmra_i5g22:31518061
7df3h592:31512476
cp_cls_t_1082:31535311
logging_enabled_new:31498466
j0d79568:31499440
jb_cp_cls_t_632:31543129
b1ei0813:31506238
748c7209:31512887
32d76977:31512328
ha629193:31508444
cdk-lw-on:31524445
f9bf7985:31543584
2cfga817_c:31522397
a1ije391_t:31540920
api_cot_ctrl:31513533
g4a0e106:31529594
41836166:31537719
prpt_ctrl:31518467
jbcp_cls_pctr_t:31531130
61138546:31518536
cp_intellij_c_nes:31533503
d7b18187:31526828
nes-curr-leftover-1500:31538366
h561g132:31533256
sessions_5:31532894
ihg5j128:31534457
f412h606:31546039
9gg7f176:31542112
nes-nc-nf-2k:31546582
```
Contributor guide
Assessment
This issue has not been assessed yet.