Livekit inference GPT-5 mini does not works

Getting Started
,
1

Livekit inference model GPT-5 mini mentioned in openai section does not work

2

I just tested with:

llm=inference.LLM(model="openai/gpt-5-mini"),

Run with uv run src/agent.py console and it seems to work. Am I missing something in my reproduction?

3

It does not work when extra_kwargs is passed without reasoning_effort

// not works

return inference.LLM(

model=f"openai/gpt-5-mini",

extra_kwargs={

               "max_completion_tokens": 250

            },

)

// works

return inference.LLM(

model=f"openai/gpt-5-mini",

extra_kwargs={

               "max_completion_tokens": 250,

               "reasoning_effort": "none"

            },

)
4

It works, I just tested

5

+1 on Mahimai, but I did double-check the agent starter again with extra_kwargs but without reasoning_effort, and it still works for me.

Can I please ask you to test with agent-starter-python/src/agent.py at main · livekit-examples/agent-starter-python · GitHub but replacing the llm portion as follows:

            # See all available models at https://docs.livekit.io/agents/models/llm/
            # llm=inference.LLM(model="openai/gpt-5.2-chat-latest"),
            llm=inference.LLM(model="openai/gpt-5-mini",
                extra_kwargs={
                    "max_completion_tokens": 250
                }),

Then we are at least testing with the same minimum reproducible example.

6

We can mark as resolved, thanks. it works in starter but somehow not in mine, i will check and see why