Paper 2601.04720
Qwen3-VL-Embedding and Qwen3-VL-Reranker: A Unified Framework for State-of-the-Art Multimodal Retrieval and Ranking
- Published
- Jan 2026
- Research lab
- Qwen
- Citations
- 167
- GitHub
- 1.3K stars
01 In brief
Summary
The report introduces Qwen3-VL-Embedding and Qwen3-VL-Reranker, a unified framework for multimodal retrieval built on the Qwen3-VL foundation model.
The embedding model uses a multi-stage training pipeline (contrastive pre-training, multi-task contrastive learning, and reranker distillation) to produce high-dimensional vectors, supporting Matryoshka Representation Learning and quantization-aware training.
The reranker uses a cross-encoder architecture for fine-grained relevance scoring.
Both models support over 30 languages and are available in 2B and 8B parameter sizes.
On MMEB-V2, Qwen3-VL-Embedding-8B achieves a state-of-the-art score of 77.8, surpassing all baselines.
On MTEB Multilingual, it scores 67.9.
The reranker models also show competitive performance, with the 8B variant improving ranking by 4.1 points over the 2B.
The report details the architecture, training objectives, data synthesis, and evaluation, demonstrating effectiveness across image-text retrieval, visual question answering, and video-text matching.
02 From the paper
Abstract
In this report, we introduce the Qwen3-VL-Embedding and Qwen3-VL-Reranker model series, the latest extensions of the Qwen family built on the Qwen3-VL foundation model. Together, they provide an end-to-end pipeline for high-precision multimodal search by mapping diverse modalities, including text, images, document images, and video, into a unified representation space. The Qwen3-VL-Embedding model employs a multi-stage training paradigm, progressing from large-scale contrastive pre-training to reranking model distillation, to generate semantically rich high-dimensional vectors. It supports Matryoshka Representation Learning, enabling flexible embedding dimensions, and handles inputs up to 32k tokens. Complementing this, Qwen3-VL-Reranker performs fine-grained relevance estimation for query-document pairs using a cross-encoder architecture with cross-attention mechanisms. Both model series inherit the multilingual capabilities of Qwen3-VL, supporting more than 30 languages, and are released in $\textbf{2B}$ and $\textbf{8B}$ parameter sizes to accommodate diverse deployment requirements. Empirical evaluations demonstrate that the Qwen3-VL-Embedding series achieves state-of-the-art results across diverse multimodal embedding evaluation benchmarks. Specifically, Qwen3-VL-Embedding-8B attains an overall score of $\textbf{77.8}$ on MMEB-V2, ranking first among all models (as of January 8, 2025). This report presents the architecture, training methodology, and practical capabilities of the series, demonstrating their effectiveness on various multimodal retrieval tasks, including image-text retrieval, visual question answering, and video-text matching.