Reasoning using LiteLLM configured with VertexAI Claude 3.7

[elvis-cai]

I just saw Librechat support Claude 3.7 with thinking(#6008), and that is awesome!
since we are using Litellm as AI gateway configured with VertexAI claude 3.7, we are curious if we could use it with thinking mode enabled as well, thank you

This is a working example of using OpenAI SDK with Litellm vertexai claude 3.7 configured

from openai import OpenAI

client = OpenAI(
  base_url="https://litellm.abc.com",
  api_key="sk-xxxx"
)

response = client.chat.completions.create(
    model="vertex_ai/claude-3-7-sonnet@20250219",
    stream=True,
    messages=[
        {"role": "system", "content": "You are a helpful assistant designed to output JSON."},
        {"role": "user", "content": "Melbourne best coffee shop"}
    ],
    extra_body={
        "thinking": {
            "type": "enabled",
            "budget_tokens": 1024
        }
    }
)


print(f"response: {response}")

Let me know if you need more info, thanks.