Qwen
/

Reasoning or Non-reasoning model?

#6
by dipta007 - opened

I am confused, is it a reasoning model or non-reasoning?

To be more specific, is it a successor of Qwen2.5 or QwQ, or is it unified?

This is a unified reasoning/non-reasoning model. You can control it at inference time.

So can I safely assume in most of the cases Qwen3-32b is better than QwQ32b?
Reasoning tasks?

https://qwenlm.github.io/blog/qwen3/

I guess you can choose.
To disable thinking, you just need to make changes to the argument enable_thinking like the following:

text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True,
    enable_thinking=False  # True is the default value for enable_thinking.
)

For deployment, you can use sglang>=0.4.6.post1 or vllm>=0.8.4 to create an OpenAI-compatible API endpoint:

Screenshot 2025-04-29 at 10.04.02 AM.png

So can I safely assume in most of the cases Qwen3-32b is better than QwQ32b?
Reasoning tasks?

Yep, I think so

Your need to confirm your account before you can post a new comment.

Sign up or log in to comment