A new benchmark for multimodal retrieval: Qwen3-VL-Embedding & Reranker is open source!

The Tongyi Qianwen team has officially open-sourced its multimodal retrieval models Qwen3-VL-Embedding and Qwen3-VL-Reranker. These models, built on Qwen3-VL, are specifically designed for understanding and retrieving mixed content such as text, images, and videos, and can handle multiple modal inputs, including text, images, visual documents, and videos, within a unified framework.