Responses API (Reasoning Models)

Use the responses API to query o-series and similar models that provide reasoning efforts.

1
from meshapi import ResponsesParams
2
3
reply = client.responses.create(
4
    ResponsesParams(
5
        model="openai/o4-mini",
6
        input="Explain the halting problem in two sentences.",
7
        reasoning={"effort": "medium"},
8
        max_output_tokens=512,
9
    )
10
)
11
12
print(reply.output)

Streaming works the same way via client.responses.stream(params).

Use client.responses.list() and client.responses.get("resp_...") for persisted/background response jobs. A response returned from a synchronous create call is not guaranteed to be retrievable later by ID.