# tencent/WeVisOne-VL-4B · Hugging Face — retained record

> Retained reddit record from reddit:LocalLLaMA.

Canonical: https://highsignal.app/data/records/7ccb0fcfd96ab819

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


source: [reddit](https://highsignal.app/data/reddit) (reddit:LocalLLaMA)  published 2026-10-23 08:14 ingested 2026-10-24 02:40

[original source ↗](https://www.reddit.com/r/LocalLLaMA/comments/1xd85ki/tencentwevisonevl4b_hugging_face/)

## retained text



```
[https://huggingface.co/tencent/WeVisOne-VL-2B](https://huggingface.co/tencent/WeVisOne-VL-2B)

# WeVisOne-VL: Unified Visual Perception from Objects to Pixels

**WeVisOne-VL** (**We**Chat **Vis**ion All-in-**One**) is a unified vision-language model developed by WeChat Vision, Tencent Inc. Built upon Qwen3-VL, it unifies object-level perception (open-vocabulary object detection, referring localization, object counting, object pointing, and OCR) and pixel-level perception (referring segmentation and promptable alpha matting) through structured next-token prediction and a lightweight 9.6M-parameter pixel decoder.


```



document fetched 2026-10-24 00:42 · canonical https://reddit.com/r/LocalLLaMA/comments/1xd85ki/tencentwevisonevl4b_hugging_face
