Qwen

/

Qwen3-235B-A22B-Instruct-2507

like 650

Follow

Qwen 45.9k

We have encountered often the following response:

The request params:
"max_tokens": 32768,
"temperature": 0.1,
"tool_choice": "auto"

{
  "id": "chatcmpl-9d0b844f4ae34a73a6c8749e3ec3b7ae",
  "model": "Qwen3-235B-A22B-Instruct-2507",
  "usage": {
    "total_tokens": 61340,
    "prompt_tokens": 61325,
    "completion_tokens": 15,
    "prompt_tokens_details": null,
    "completion_tokens_details": {
      "audio_tokens": null,
      "reasoning_tokens": 0,
      "accepted_prediction_tokens": null,
      "rejected_prediction_tokens": null
    }
  },
  "object": "chat.completion",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "Now I'll check the build script for WeixinMinigame support:\n\n",
        "tool_calls": null,
        "function_call": null
      },
      "finish_reason": "stop"
    }
  ],
  "created": 1753952013,
  "system_fingerprint": null
}

Apparently from the choices content it should choose a tool and finish with "tool_calls" but the above reply would make the workflow stop unexpectedly.

Is there any one having such issue as well? What is the suggested way/prompts to avoid such issue?