[Usage]: MCP-USE with VLLM gpt-oss:20b via ChatOpenAI

[Tahirc1]

Your current environment

The output of `python collect_env.py`

How would you like to use vllm

i am trying to create an agent using gpt-oss:20B with mcp-use

most times the model returns "Agent completed the task successfully.", and sometimes the proper output which is required

code

vllm serve openai/gpt-oss-20b --max-model-len 100000 --gpu-memory-utilization 0.9 --port 8000 --tool-call-parser openai --enable-auto-tool-choice

client = MCPClient.from_dict(config)
llm = ChatOpenAI(
model="openai/gpt-oss-20b",
base_url="http://127.0.0.1:8000/v1",
api_key="not-needed",
temperature=0.8,
max_tokens=2048
)
agent = MCPAgent(llm=llm, client=client, max_steps=30)

also raising this on mcp-use

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.