QwenLM/Qwen3-VL

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

19847 stars · 1841 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

Qwen3-VL is a series of open-weight multimodal vision-language models from Alibaba's Qwen team, available in Dense and MoE architectures with Instruct and Thinking editions. The repository provides model weights, usage examples, cookbooks, and tooling for running the models for image/video understanding, OCR, visual coding, and GUI agent tasks.

Use cases

• understand images and videos with a vision-language model
• extract text from images with multilingual OCR
• build a GUI agent that operates PC or mobile apps visually
• generate HTML/CSS/JS or Draw.io diagrams from screenshots
• do 2D and 3D visual grounding and spatial reasoning
• answer math and STEM questions with multimodal reasoning
• run long-context document and hours-long video analysis

When to choose

• you need an open-weight multimodal LLM you can self-host or fine-tune
• your app combines text, image, and video understanding in one model
• you need strong OCR across many languages and degraded image conditions
• you want visual agent capabilities like GUI element recognition and tool use

When to avoid

• you only need text-only LLM inference without vision
• you lack GPU resources and cannot use an API instead
• you need a small edge model outside the offered size range

Facets

library · maturity active

machine-learning deep-learning llm-inference ocr computer-vision agent-framework large-language-models computer-vision artificial-intelligence python cross-platform vision-language-model multimodal qwen video-understanding visual-grounding moe open-weights natural-language-processing ai-agents gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="QwenLM/Qwen3-VL")