Back to events
Model releaseHigh global significanceConfirmed confidence

Alibaba Cloud releases Qwen2.5-VL

Alibaba Cloud’s Qwen team released the Qwen2.5-VL vision-language model family in 3B, 7B, and 72B sizes, publishing both base and instruction-tuned models.

Event details

The Qwen team released Qwen2.5-VL on January 26, 2025. The family accepts text, image, and video inputs and generates text. Base and instruction-tuned models in 3B, 7B, and 72B sizes were made available through Hugging Face and ModelScope.

Why it matters

Qwen2.5-VL extended image, video, document, and visual-agent capabilities while offering publicly downloadable weights at several parameter scales.

79/100Global significance score. Regional effects are recorded only when the evidence supports a meaningful difference.

Access notes

Model weights are public under model-specific licenses; the 72B instruction model is also available through Qwen Chat.