Cannot use custom model added

Support Bug Reports
1

Where does the bug appear (feature/product)?

Cursor IDE

Describe the Bug

Keep getting error “Model claude-opus-4-5-20251101 is not a reasoning effort model”

I added custom model “claude-opus-4-5-20251101”

Steps to Reproduce

Go to Settings → Models → Add custom model “claude-opus-4-5-20251101”
Enable Anthropic API Key

In chat, try to use that model.

Screenshots / Screen Recordings

Screenshot 2025-11-30 at 11.12.48 PM.png
Screenshot 2025-11-30 at 11.12.48 PM.png1542×1876 112 KB

Operating System

MacOS

Current Cursor Version (Menu → About Cursor → Copy)

Version: 2.1.39
VSCode Version: 1.105.1
Commit: 60d42bed27e5775c43ec0428d8c653c49e58e260
Date: 2025-11-27T02:30:49.286Z
Electron: 37.7.0
Chromium: 138.0.7204.251
Node.js: 22.20.0
V8: 13.8.258.32-electron.0
OS: Darwin arm64 25.1.0

For AI issues: which model did you use?

claude-opus-4-5-20251101

For AI issues: add Request ID with privacy disabled

Request ID: 723cb093-da20-4d12-9231-9eee7c8d7998
{“error”:“ERROR_OPENAI”,“details”:{“title”:“Unable to reach the model provider”,“detail”:“We encountered an issue when using your API key: Model claude-opus-4-5-20251101 is not a reasoning effort model\n\nAPI Error:\n\n\nModel claude-opus-4-5-20251101 is not a reasoning effort model\n”,“additionalInfo”:{},“buttons”:,“planChoices”:},“isExpected”:true}
ConnectError: [invalid_argument] Error
at UZc.$endAiConnectTransportReportError (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.desktop.main.js:6331:408701)
at rLo._doInvokeHandler (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.desktop.main.js:6948:21873)
at rLo._invokeHandler (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.desktop.main.js:6948:21615)
at rLo._receiveRequest (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.desktop.main.js:6948:20377)
at rLo._receiveOneMessage (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.desktop.main.js:6948:19194)
at _Pt.value (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.desktop.main.js:6948:17286)
at _e._deliver (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.desktop.main.js:49:2962)
at _e.fire (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.desktop.main.js:49:3283)
at tpt.fire (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.desktop.main.js:6316:12156)
at MessagePort. (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.desktop.main.js:8976:18439)

Does this stop you from using Cursor

Yes - Cursor is unusable

Custom Models not visible and cannot delete
2

hi @user1324 are you using a Bedrock API or a custom key?

Its most likely duplicate of if you either use Bedrock or a custom key:

3

@condor No, I am just using anthropic custom key.

Screenshot 2025-12-01 at 10.21.41 AM
Screenshot 2025-12-01 at 10.21.41 AM1446×1948 235 KB

4

@condor can you please help?

5

@deanrie could you please help me with this? This is a long pending issue and never been solved.

6

Hey, thanks for your patience.

You don’t need to add a custom model named “claude-opus-4-5-20251101”. Instead:

  1. Go to Cursor Settings > Models > Model Names
  2. Enable your Anthropic API key
  3. Pick the built-in “Opus 4.5” model from the chat dropdown
  4. All requests will automatically go through your API key

So you should use the models that are already in Cursor, not create a custom one. Your API key will be used automatically to pay for requests.

Try this and let me know if it worked.

7

@deanrie, I see error while trying to use Opus 4.5 with Anthropic Key. I tried with multiple API keys

Request ID: 9b710733-7a45-4839-b3d1-dc869f3739ae
{“error”:“ERROR_OPENAI”,“details”:{“title”:“Unable to reach the model provider”,“detail”:“We encountered an issue when using your API key: Streaming error\n\nAPI Error:\n\n\n{\"error\":{\"type\":\"provider\",\"reason\":\"provider_error\",\"message\":\"Provider returned 401\",\"retryable\":false,\"provider\":{\"status\":401,\"body\":\"{\\\"type\\\":\\\"error\\\",\\\"error\\\":{\\\"type\\\":\\\"authentication_error\\\",\\\"message\\\":\\\"invalid x-api-key\\\"},\\\"request_id\\\":\\\"req_011CWixcmFxnnrnYrKfbDLf1\\\"}\"}}}\n”,“additionalInfo”:{},“buttons”:,“planChoices”:},“isExpected”:true}

8

I see that after switching to the built-in model, a different error showed up.

The “invalid x-api-key” error is often related to the “Override OpenAI Base URL” setting. Can you check:

  • Settings > Models > API Keys
  • Make sure “Override OpenAI Base URL” is turned off
  • If it was on, turn it off and restart Cursor
  • Open a new chat and try again

If that doesn’t help, please send:

  • A screenshot of Settings > Models > API Keys (with all toggles visible)
  • Your current Cursor version (Help > About Cursor)
  • Confirmation that Override OpenAI Base URL is definitely turned off

This is a known issue, I’ll add your case to the ticket for the dev team.

9

@deanrie

image
image3420×2214 501 KB

Version: 2.2.44
VSCode Version: 1.105.1
Commit: 20adc1003928b0f1b99305dbaf845656ff81f5d0
Date: 2025-12-24T21:41:47.598Z
Electron: 37.7.0
Chromium: 138.0.7204.251
Node.js: 22.20.0
V8: 13.8.258.32-electron.0
OS: Darwin arm64 25.2.0

Override OpenAI Base URL was never turned ON.

12

Thanks for the info.

We checked the logs, and the 401 error is coming directly from the Anthropic API, which means Anthropic isn’t accepting the API key itself.

Please check:

  • Your Anthropic API key starts with sk-ant-api... (not just sk-ant-)
  • The key is active in your Anthropic Console
  • After changing or re-entering the key in Cursor, start a new chat

To confirm the key works outside Cursor, try this in your terminal:

curl https://api.anthropic.com/v1/messages \
  -H "x-api-key: YOUR_KEY_HERE" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{"model":"claude-opus-4-5-20251101","max_tokens":10,"messages":[{"role":"user","content":"Hey"}]}'

If you get a response, the key works and the issue is on our side. If you get a 401, the issue is with the key.

Let me know what you get.

1 Like
13

This topic was automatically closed 22 days after the last reply. New replies are no longer allowed.