FreeLLMTencent releases WeVisOne-VL, a unified visual-language model for object and pixel perception
WeVisOne-VL unifies object-level and pixel-level visual tasks through next-token prediction, with code and model collection available on GitHub and Hugging Face.
Tencent's research team introduced WeVisOne-VL, short for WeChat Vision All-in-One, as a new vision-language model.
The model aims to handle both object-level and pixel-level perception by predicting structured tokens in a unified framework.
Details of the approach are described in the paper titled "WeVisOne-VL: Unified Visual Perception from Objects to Pixels".
The project's source code is hosted on GitHub at https://github.com/Tencent/WeVisOne-VL.
Model checkpoints and related resources are available through a Hugging Face collection at https://huggingface.co/collections/tencent/wevisone-vl.
What it means for you
WeVisOne-VL offers researchers a single model that can address diverse visual tasks, reducing the need for separate specialized models. For developers seeking free or locally runnable AI, the open-source release provides immediate access to a versatile visual-language system.
Sources
- Tencent paper: WeVisOne-VL: Unified Visual Perception from Objects to Pixels, Hugging Face Papers (official), 21 Oct 2026
Written automatically from the sources above, then checked by our code: every number and name must appear in them, and no sentence may copy a source. Spotted a mistake? Report it.
Which AI can you use for free today?
Every free AI model, ranked by blind human votes, with the memory and speed it needs on your own GPU.