docker / docker/model-runner

Disable thinking if reasoning_effort is none.

Open
#1,028 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Go
Stars
651
Forks
155
PR merge metrics
No merged PRs in 30d

Description

To maintain the open AI compatibility of your API, would it be possible to disable the reasoning of the LLM if reasoning_effort: "none" is specified in the request?

I know the standard way to disable the thinking for models such as Gemma 4 and Qwen 3 would be { ..., chat_template_kwargs: {"enable_thinking": false} }, but I don't want to have to write VLLM/LLAMA specefic parameters in OpenAI request.

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by tracing the OpenAI-compatible request handling and the existing mapping for chat_template_kwargs or reasoning_effort. Confirm how reasoning_effort: "none" is represented for supported models, then verify that it disables thinking without requiring provider-specific parameters while other effort values remain unchanged.

Written by the indexing model from the issue text.

Assessment

Tech stack
go
Domain
ai, api
Issue type
Feature
Difficulty
3/5
Estimated time
1-2 days
Activity status
Quiet
Clarity
Mostly clear
Newbie friendliness
62/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.