Tencent WeChat Vision’s vision-language model unifies object detection, localization, counting, OCR, pixel-level segmentation, and matting in one system, with 2B and 4B weights.
Only published specifications are shown.
This record covers multiple tiers. Modalities and capabilities describe the group coverage; support and licensing depend on the individual tier.
— Tencent WeChat Vision authors introduced WeVisOne-VL in a research report, unifying vision-language tasks from detection, localization, and OCR through segmentation and matting. ↗ — Tencent’s official repositories released WeVisOne-VL-2B and WeVisOne-VL-4B weights under Apache-2.0. ↗
WeVisOne-VL was introduced in a research report on 2026-10-22, with 2B and 4B weights released on October 23 under Apache-2.0. Its official model card says it is built on Qwen3-VL; the archive has no matching Qwen3-VL model entity, so no unsupported lineage edge is added.
Tencent WeChat Vision authors introduced WeVisOne-VL in a research report, unifying vision-language tasks from detection, localization, and OCR through segmentation and matting.
View sourceTencent’s official repositories released WeVisOne-VL-2B and WeVisOne-VL-4B weights under Apache-2.0.
View sourceSelect a card to view the related model.
No related models recorded yet
Copy this template, add the field, proposed value, and primary source, and share it with the archive maintainer.