[Bug]: GPT-OSS Tool Answer Role Object not working

[larbikhalid]

Your current environment

docker run --runtime nvidia --gpus all \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HUGGING_FACE_HUB_TOKEN=$HF_TOKEN" \
    -p 8000:8000 \
    --ipc=host \
    **vllm/vllm-openai:v0.10.2** \
    --model openai/gpt-oss-20b
    --tool-call-parser openai \
    --enable-auto-tool-choice

🐛 Describe the bug

while this body works:

{
    "model": "openai/gpt-oss-20b",
    "messages": [
        {
            "role": "user",
            "content": [
                {
                    "text": "what is the weather in Paris",
                    "type": "text"
                }
            ]
        },
        {
            "role": "assistant",
            "tool_calls": [
                {
                    "function": {
                        "arguments": "{\"city\": \"Paris\"}",
                        "name": "get_weather"
                    },
                    "id": "chatcmpl-tool-aaae1ecc31344b9d94f69f52a0a4c52e",
                    "type": "function"
                }
            ]
        },
        {
            "role": "tool",
            "tool_call_id": "chatcmpl-tool-aaae1ecc31344b9d94f69f52a0a4c52e",
            "content": **"18 degrees"**
        }
    ],
    "tools": [
        {
            "type": "function",
            "function": {
                "name": "get_weather",
                "description": "get weather using city.",
                "parameters": {
                    "type": "object",
                    "required": []
                }
            }
        }
    ],
    "tool_choice": "auto"
}

The next one does not;

{
    "model": "openai/gpt-oss-20b",
    "messages": [
        {
            "role": "user",
            "content": [
                {
                    "text": "what is the weather in Paris",
                    "type": "text"
                }
            ]
        },
        {
            "role": "assistant",
            "tool_calls": [
                {
                    "function": {
                        "arguments": "{\"city\": \"Paris\"}",
                        "name": "get_weather"
                    },
                    "id": "chatcmpl-tool-aaae1ecc31344b9d94f69f52a0a4c52e",
                    "type": "function"
                }
            ]
        },
        {
            "role": "tool",
            "tool_call_id": "chatcmpl-tool-aaae1ecc31344b9d94f69f52a0a4c52e",
            "content": **[
                {
                    "text": "18 degrees",
                    "type": "text"
                }
            ]**
        }
    ],
    "tools": [
        {
            "type": "function",
            "function": {
                "name": "get_weather",
                "description": "get weather using city.",
                "parameters": {
                    "type": "object",
                    "required": []
                }
            }
        }
    ],
    "tool_choice": "auto"
}

the output is:

{
    "error": {
        "message": "1 validation error for Message\ncontent.0\n  Input should be a valid dictionary or instance of Content [type=model_type, input_value=[{'text': '18 degrees', 'type': 'text'}], input_type=list]\n    For further information visit https://errors.pydantic.dev/2.11/v/model_type None",
        "type": "BadRequestError",
        "param": null,
        "code": 400
    }
}

according to openai doc and the example in /docs of vllm swagger:
chat-create-messages-tool-message-content

tool message can also be an array

Before submitting a new issue...

• Make sure you already searched for relevant issues, and asked the chatbot living at the bottom right corner of the documentation page, which can answer lots of frequently asked questions.

[iwaitu]

same proplem.

docker run -it --rm \
  --gpus all \
  --ipc=host \
  -p 8000:8000 \
  -e CUDA_DEVICE_ORDER=PCI_BUS_ID \
  -v /models/gpt-oss-120b/gpt-oss-120b:/models/gpt-oss-120b:ro \
  --name gptoss \
  vllm/vllm-openai:v0.10.2 \
  --model /models/gpt-oss-120b \
  --served-model-name gpt-oss-120b \
  --tensor-parallel-size 2 \
  --tool-call-parser openai 

tools: List[Dict[str, Any]] = [
        {
            "type": "function",
            "function": {
                "name": "get_weather",
                "description": "查询指定城市当前天气",
                "parameters": {
                    "type": "object",
                    "properties": {
                        "city": {"type": "string", "description": "城市名，如：北京、上海、深圳、柏林等"}
                    },
                    "required": ["city"]
                }
            }
        }
    ]

    messages: List[Dict[str, Any]] = [
        {"role": "system", "content": "你是一个有用的助手，擅长回答与天气相关的问题。"},
        {"role": "user", "content": "请问柏林现在的天气如何？"}
    ]

    payload = {
        "model": MODEL_ID,
        "messages": messages,
        "tools": tools,
        "tool_choice": "auto",
        "max_tokens": 300,
        "temperature": 0.3
    }

got response:

{
  "id": "chatcmpl-4332321e146149d888cdd3cf46d1ddd1",
  "object": "chat.completion",
  "created": 1758483629,
  "model": "gpt-oss-120b",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "柏林当前的天气情况是：\n\n- **天气状况**：多云  \n- **气温**：约 22 °C  \n- **湿度**：约 68 %  \n- **风速**：约 13 km/h，西北风  \n- **气压**：约 1015 hPa  \n\n如需更详细的预报（如未来 24 小时或 7 天趋势），请告诉我！",
        "refusal": null,
        "annotations": null,
        "audio": null,
        "function_call": null,
        "tool_calls": [],
        "reasoning_content": "User asks current weather in Berlin. Need to call get_weather with city \"柏林\".\n{\n  \"city\": \"柏林\"\n}"
      },
      "logprobs": null,
      "finish_reason": "stop",
      "stop_reason": null,
      "token_ids": null
    }
  ],
  "service_tier": null,
  "system_fingerprint": null,
  "usage": {
    "prompt_tokens": 162,
    "total_tokens": 312,
    "completion_tokens": 150,
    "prompt_tokens_details": null
  },
  "prompt_logprobs": null,
  "prompt_token_ids": null,
  "kv_transfer_params": null
}

[luis5tb]

I'm also hitting this same issue. Working with:
{"role": "tool", "content": "462a2866-1c36-424a-a457-6b4b6e5fa27a"}
But not working with:
{"role": "tool", "content": [{ "type": "text","text": "f2e897a7-2705-4337-8193-2a8f57b81618"}]}