Qwen3-VL
Qwen3-VL is Alibaba's open vision-language model family after Qwen2.5-VL. For OCR it reads 32 languages and copes better with blur, low light and tilted text.
A release of Qwen-VL. Other versions: Qwen2.5-VL, Qwen2-VL.
- Released
- September and October 2025
- Model ID
- Qwen/Qwen3-VL-2B-Instruct, Qwen/Qwen3-VL-4B-Instruct, Qwen/Qwen3-VL-8B-Instruct, Qwen/Qwen3-VL-32B-Instruct, Qwen/Qwen3-VL-30B-A3B-Instruct, Qwen/Qwen3-VL-235B-A22B-Instruct
- Size
- Dense 2B, 4B, 8B and 32B. Mixture-of-experts 30B-A3B and 235B-A22B
- License
- Apache-2.0
- Also searched as
- qwen3-vl, qwen 3 vl, Qwen3-VL-8B-Instruct
What is new in Qwen3-VL
- OCR in 32 languages, up from 19 in Qwen2.5-VL.
- More robust in low light, blur and tilt, and with rare or ancient characters and jargon.
- Better structure parsing on long documents.
- 256K context, expandable to 1M.
- Can turn an image into Draw.io, HTML, CSS or JavaScript.
Use Qwen3-VL through an API
OCRRouter runs a different release of Qwen-VL as qwen/qwen-vl.
Qwen-VL versions
- Qwen-VL: all versions, price, speed and features on OCRRouter
- Qwen3-VL (September and October 2025), this page
- Qwen2.5-VL (January 28, 2025 (32B in March 2025))
- Qwen2-VL (August 29, 2024)
Questions
Sources
Last updated .