[BUG] Azure AI Foundry model probe missing api-version parameter causes spurious 404 error

Resolved 💬 1 comment Opened Jun 2, 2026 by XAltairX Closed Jul 6, 2026

Preflight Checklist

  • [x] I have searched existing issues and this hasn't been reported yet
  • [x] This is a single bug report (please file separate reports for different bugs)
  • [x] I am using the latest version of Claude Code

What's Wrong?

When using Claude Code with Azure AI Foundry (CLAUDE_CODE_USE_FOUNDRY=1), the startup model availability probe returns a 404 error even though actual inference works correctly. A persistent banner is shown: "Your gateway couldn't serve claude-opus-4-6".

The probe sends a request to /models/chat/completions without the required api-version query parameter:

requestUrl: https://<resource>.services.ai.azure.com/models/chat/completions
Azure AI Foundry requires api-version on every request:

https://<resource>.services.ai.azure.com/models/chat/completions?api-version=2024-12-01-preview
The actual inference SDK includes this parameter, which is why chat works despite the probe failing.

What Should Happen?

The model probe should include the api-version query parameter when hitting Azure AI Foundry endpoints, matching what the inference SDK does.
Alternatively, provide a setting to suppress or skip the model availability probe (e.g., CLAUDE_CODE_SKIP_MODEL_PROBE=1).

Error Messages/Logs

message: Azure AI Foundry rejected model "claude-opus-4-6" (HTTP 404)
httpStatus: 404
requestUrl: https://<resource>.services.ai.azure.com/models/chat/completions
probedModel: claude-opus-4-6
responseBody: {"error":{"code":"api_not_supported","message":"Requested API is currently not supported","details":"Requested API is currently not supported"}}
endpoint: https://<resource>.services.ai.azure.com/

Steps to Reproduce

Configure Claude Code with Azure AI Foundry:
CLAUDE_CODE_USE_FOUNDRY=1
ANTHROPIC_FOUNDRY_RESOURCE=<resource-name>
ANTHROPIC_FOUNDRY_API_KEY=<key>
Set model to claude-opus-4-6
Start Claude Code
Observe the gateway error banner, even though chat/inference works normally

Claude Model

Not sure / Multiple models

Is this a regression?

I don't know

Last Working Version

_No response_

Claude Code Version

2.1.156

Platform

Other

Operating System

Windows

Terminal/Shell

PowerShell

Additional Information

The error is purely cosmetic, inference works correctly. There is no setting, flag, or environment variable to suppress the probe error.

View original on GitHub ↗

This issue has 1 comment on GitHub. Read the full discussion on GitHub ↗