# LessThink-Qwen3-4B: the same model, with far less thinking [P] — retained record

> Retained reddit record from reddit:MachineLearning.

Canonical: https://highsignal.app/data/records/f0be9a0562f1d4a8

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:MachineLearning)  published 2026-09-30 07:19 ingested 2026-10-01 02:40

[original source ↗](https://www.reddit.com/r/MachineLearning/comments/1wtygav/lessthinkqwen34b_the_same_model_with_far_less/)

## retained text



```
I post-trained Qwen3-4B to spend 44% fewer tokens on reasoning, keeping its knowledge and answer style. The whole pipeline ran on one GPU.

folks, you can check it out on : https://5ivatej.com/lessthink/
```



document fetched 2026-10-01 00:30 · canonical https://reddit.com/r/MachineLearning/comments/1wtygav/lessthinkqwen34b_the_same_model_with_far_less
