[Feature][Responses API] Stream Function Call

[simon-mo]

🚀 The feature, motivation and pitch

External function call invocation is not streamed right now.

This functions do not have ways to handle it

vllm/vllm/entrypoints/openai/serving_responses.py

Line 781 in c32e6ad

async def responses_stream_generator(

and therefore the function call isn't streamed to users.

Alternatives

No response

Additional context

No response

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.

[chaunceyjiang]

Please assign it to me.
I have already implemented non-streaming function calls.

#20874

[chuddlestonCBANC]

@chaunceyjiang Hi there, is there a timeline on when this will be resolved?

[PedroMiolaSilva]

Hey guys, any news here?

[levunet]

@PedroMiolaSilva
It seems to be resolved, but the PR is still pending. (#24954)

[chaunceyjiang]

Most of the work has already been done, and it’s now waiting for review.
#24317

[qandrew]

i've also added a couple of PRs (#24759, #24556, etc) that implement streaming for GPT-OSS

[bartlettroscoe]

Any recent PRs being worked on to address streaming tool calling response API?