Thanks for the reply. I tried this

import requests

# Update Tool (PATCH /tool/:id)
response = requests.patch(
  "<MY_URL>",
  headers={
    "Authorization": "Bearer <MY_BEARER>",
    "Content-Type": "application/json"
  },
  json={
    "async": false,
    "messages": [],
    "function": {
      "name": "<MY_NAME>",
      "description": "<MY_DESCRIPTION>",
      "strict": true,
      "parameters": {
        "type": "object",
        "properties": {
          "vector_inputs": {
            "type": "array",
            "description": "<MY_DESCRIPTION>"
          }
        },
        "required": [
          "vector_inputs"
        ]
      }
    },
    "server": {
      "url": "<MY_URL>",
      "timeoutSeconds": 120
    }
  },
)

print(response.json())

but I still see this error in the logs

14:02:47:786
[LOG]
Model called tool: <TOOL_CALL>

14:02:47:801
[LOG]
Model request cost (attempt #1, $0.02058, 4116 prompt, 0 completion)

14:02:47:806
[LOG]
Voice input: Just a sec

14:02:48:609
[INFO]
Turn latency: 2115ms (transcriber: 411ms, endpointing: 102ms, model: 0ms, voice: 0ms)

14:02:48:609
[CHECKPOINT]
Assistant speech started

14:02:49:663
[CHECKPOINT]
Assistant speech stopped

14:02:49:664
[LOG]
Endpointing timeout 100ms (rule: `heuristic`)

14:02:49:767
[LOG]
Model request started (attempt #1, gpt-4o-2024-05-13, azure-openai, westus3)

SO it still seems the timeout is 100 ms