An error occurred when the agents used the Claude model.

[Passerby1011]

What happened?

When I encountered an error while using the claude-3-7-sonnet-20250219 model in agents.
I have turned off the switch of thinking.

Version Information

root@home:/home/home/LibreChat# docker images | grep librechat
ghcr.io/danny-avila/librechat-dev latest 59660cbf7aa2 22 hours ago 872MB

Steps to Reproduce

Use the claude-3-7-sonnet-20250219 model in the agent.

What browsers are you seeing the problem on?

No response

Relevant log output

{"error":{"message":"thinking.disabled.budget_tokens: Extra inputs are not permitted  (request id: 2025030823182553645743AI8e20Ww)","type":"invalid_request_error"}} 
---
2025-03-08T15:26:22.956Z debug: [saveConvo] api/app/clients/BaseClient.js - saveMessageToDatabase #saveConvo

2025-03-08T15:27:04.083Z error: [api/server/controllers/agents/client.js #sendCompletion] Operation aborted 400 {"type":"bad_response_status_code","error":{"type":"upstream_error","m... [truncated]

2025-03-08T15:27:04.084Z error: [api/server/controllers/agents/client.js #sendCompletion] Unhandled error type 400 {"type":"bad_response_status_code","error":{"type":"upstream_error"... [truncated]

2025-03-08T15:27:04.086Z error: [handleAbortError] AI response error; aborting request: 400 {"type":"bad_response_status_code","error":{"type":"upstream_error","message":"Provider AP... [truncated]

2025-03-08T15:27:04.116Z debug: [AgentController] Request closed

Screenshots

Code of Conduct

• I agree to follow this project's Code of Conduct

[danny-avila]

Can't be reproduced with/without thinking enabled

[Passerby1011]

I don't know why that's the case, either. I'm normal in normal conversations.

[Passerby1011]

The main errors are {"error":{"message":"thinking.disabled.budget_tokens:Extra inputs are not permitted (request id:2025030823182553645743AI8e20Ww)","type":"invalid_request_error"}}

[Shinobu-Kazahana]

Im getting the same issue

I cannot disable thinking, Get the error each time.

edit:
settings