Qwen3-VL Technical Report

By Shuai Bai, Yuxuan Cai, Ruizhe Chen, Ke-qin Chen, Xiong-Hui Chen, Zesen Cheng, Lianghao Deng, Wei Ding, Rongyao Fang, Chang Gao and 52 more. arXiv.org, 2025

1,662 citations recorded by the source corpus.

We introduce Qwen3-VL, the most capable vision-language model in the Qwen series to date, achieving superior performance across a broad range of multimodal benchmarks. It natively supports interleaved contexts of up to 256K tokens, seamlessly integrating text, images, and video. The model family includes both dense (2B/4B/8B/32B) and mixture-of-experts (30B-A3B/235B-A22B) variants to accommodate diverse latency-quality trade-offs. Qwen3-VL delivers three core pillars: (i) markedly stronger pure-text understanding, surpassing comparable text-only backbones in several cases; (ii) robust long-context comprehension with a native 256K-token window for both text and interleaved multimodal inputs, enabling faithful retention, retrieval, and cross-referencing across long documents and videos; and (iii) advanced multimodal reasoning across single-image, multi-image, and video tasks, demonstrating leading performance on comprehensive evaluations such as MMMU and visual-math benchmarks (e.g., MathVista and MathVision). Architecturally, we introduce three key upgrades: (i) an enhanced interleaved-MRoPE for stronger spatial-temporal modeling across images and video; (ii) DeepStack integration, which effectively leverages multi-level ViT features to tighten vision-language alignment; and (iii) text-based time alignment for video, evolving from T-RoPE to explicit textual timestamp alignment for more precise temporal grounding. Under comparable token budgets and latency constraints, Qwen3-VL achieves superior performance in both dense and Mixture-of-Experts (MoE) architectures. We envision Qwen3-VL serving as a foundational engine for image-grounded reasoning, agentic decision-making, and multimodal code intelligence in real-world workflows.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/15538df854dd33351dfb5cefd7e8f3340c8936c3

Citations in the corpus

• Constrained Dominant Sets for Multimodal Document Question Answering (2026, unknown): “Unless noted, CDS uses an evidence budget B =10 , no query-row sparsification, and uniform initialization; the reader is Qwen3-VL-32B [47], with Qwen2.”
• Min-Cost Flow Routing for Evidence Assembly in Long Multimodal Documents (2026, unknown): “Qwen3-VL-32B-Instruct reads evidence and combines findings into answers, while GPT-4o-mini serves as the evaluation judge, following prior work.”