---
format: "aidr-story-markdown/v1"
id: "0f4bc2a13348e254780cb3a2528c1382e8ccb9d8fdc3a1fca37f85bb2d355c21"
canonical_url: "https://aidr.today/0f4bc2a1?lang=en"
title: "WeVisOne-VL: Unified Visual Perception from Objects to Pixels"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-10-23T00:00:00.000Z"
category: "Open Source"
topics: ["llm","open-source"]
source_urls: ["https://huggingface.co/papers/2610.31329"]
summary: "WeVisOne-VL is a vision-language model that unifies object-level and pixel-level perception through structured next-token prediction. It supports tasks such as object detection, segmentation, OCR, and matting, and uses a lightweight pixel decoder with a special token to generate instance-level pixel outputs. The model is trained with a three-stage curriculum SFT followed by mixed-task RL and reportedly achieves leading performance across over 30 benchmarks."
---

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

> [Open the canonical story](<https://aidr.today/0f4bc2a1?lang=en>)

**Published:** 2026-10-23T00:00:00.000Z
**Category:** Open Source
**Topics:** llm, open\-source

## Summary

WeVisOne\-VL is a vision\-language model that unifies object\-level and pixel\-level perception through structured next\-token prediction\. It supports tasks such as object detection, segmentation, OCR, and matting, and uses a lightweight pixel decoder with a special token to generate instance\-level pixel outputs\. The model is trained with a three\-stage curriculum SFT followed by mixed\-task RL and reportedly achieves leading performance across over 30 benchmarks\.

## Sources

- [Story source](<https://huggingface.co/papers/2610.31329>)