Custom Model Provider Error

[zhipenglv11]

Where does the bug appear (feature/product)?

Cursor IDE

Describe the Bug

PS：I wrote a Python script to test the baseUrl and apiKey, and it works fine.

Request ID: 786d6c1a-fa9f-4b51-a5b0-507ea308e902
{“error”:“ERROR_PROVIDER_ERROR”,“details”:{“title”:“Provider Error”,“detail”:“Provider returned error: {"error":{"message":"","type":""}}”,“isRetryable”:false,“additionalInfo”:{},“buttons”:,“planChoices”:},“isExpected”:true}
Provider Error Provider returned error: {“error”:{“message”:“”,“type”:“”}}
IAi: Provider Error Provider returned error: {“error”:{“message”:“”,“type”:“”}}
at c5A (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.desktop.main.js:34247:23670)
at o5A (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.desktop.main.js:34247:22658)
at g5A (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.desktop.main.js:34248:6285)
at xou.run (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.desktop.main.js:34248:10400)
at async COa.runAgentLoop (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.desktop.main.js:46811:10024)
at async F3u.streamFromAgentBackend (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.desktop.main.js:46865:9277)
at async F3u.getAgentStreamResponse (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.desktop.main.js:46865:13663)
at async bMe.submitChatMaybeAbortCurrent (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.desktop.main.js:34308:17597)
at async Ea (vscode-file://vscode-app/Applications/Cursor.app/Contents/Resources/app/out/vs/workbench/workbench.desktop.main.js:45790:4826)

Steps to Reproduce

choose my custom model

Screenshots / Screen Recordings

Operating System

MacOS

Version Information

Version: 2.6.13
VSCode Version: 1.105.1
Commit: 60faf7b51077ed1df1db718157bbfed740d2e160
Date: 2026-03-06T06:17:49.499Z
Build Type: Stable
Release Track: Default
Electron: 39.6.0
Chromium: 142.0.7444.265
Node.js: 22.22.0
V8: 14.2.231.22-electron.0
OS: Darwin arm64 23.2.0

Does this stop you from using Cursor

Yes - Cursor is unusable

[deanrie]

Hey, this is a known issue with Override OpenAI Base URL.

What’s happening: in Agent mode, Cursor is currently sending requests in the OpenAI Responses API format (input, flat tools format, etc.) instead of the standard Chat Completions format (messages, nested tools format). If your provider (api.gptsapi.net) only supports /v1/chat/completions, the request will fail. The empty error {"error":{"message":"","type":""}} points to this. The provider doesn’t understand the format and returns a vague response.

What you can try:

1. Switch from Agent mode to Ask mode. It uses the standard Chat Completions format and is more likely to work.
2. Go to Cursor Settings > Network > HTTP Compatibility Mode and switch to HTTP/1.1. Custom endpoints often don’t work well with HTTP/2.

More details: Cursor Agent sends Responses API format to /chat/completions endpoint.

The team is aware of this issue. There’s no timeline yet, but reports like this help with prioritization. Let me know if Ask mode helps.