GPT 5.6 Error when trying to prompt in IDE or Agent Window

Where does the bug appear (feature/product)?

Cursor IDE

Describe the Bug

I get this error when trying to use GPT 5.6: not sure what it means and how to fix it.

Provider returned error: {“detail”:“Validation Error”,“errors”:[{“loc”:“body → tools → 7 → type”,“msg”:“Input should be ‘function’”,“type”:“literal_error”},{“loc”:“body → tools → 7 → function”,“msg”:“Field required”,“type”:“missing”}]} Request ID: 4baca100-addf-4c67-927d-bee75b8cffc3

Steps to Reproduce

I chose the GPT5.6 and started prompting and it fails with the error/

Operating System

Windows 10/11

Version Information

Version: 3.13.10 (user setup)
VS Code Extension API: 1.128.0
Commit: 4f02290ccd9304f0e6bf8ee85f6e9106f02ac1f0
Date: 2026-07-23T21:41:07.333Z
Layout: Agent Window
Build Type: Stable
Release Track: Default
Electron: 40.10.3
Chromium: 144.0.7559.236
Node.js: 24.15.0
V8: 14.4.258.32-electron.0
xterm.js: 6.1.0-beta.256
OS: Windows_NT x64 10.0.26200

Does this stop you from using Cursor

No - Cursor works, but with this issue

Hey there! I checked the logs for your Request ID, and this isn’t a Cursor-side problem with GPT-5.6 itself. It’s coming from a custom OpenAI Base URL override in your setup.

In Cursor Settings > Models, under the OpenAI provider, there’s a custom Base URL enabled that points to a third-party endpoint (tinker.thinkingmachines.dev). Because of that, every GPT/OpenAI model you pick, including GPT-5.6 Sol, gets routed to that endpoint instead of Cursor’s own infrastructure, and that endpoint rejects the request with the validation error you’re seeing:

"body -> tools -> 7 -> type", "Input should be 'function'"

That’s also why your other models (Claude, Gemini, Composer, etc.) keep working. The override only applies to OpenAI/GPT models.

To use Cursor’s native GPT-5.6, turn off the custom OpenAI Base URL override (or remove that OpenAI key) under Cursor Settings > Models, and GPT-5.6 will route through Cursor and work normally. If you set that endpoint up on purpose for a custom model, you can just leave the override off when you want Cursor’s built-in GPT models and turn it back on for your own.

More on API keys / BYOK here: Bring your own API key | Cursor Docs

Let me know if it’s still failing after that!