Qwen2.5-VL 72B

Qwen2.5-VL 72B

Quick answer: Qwen2.5-VL 72B is Alibaba's flagship open-weight multimodal model, outperforming GPT-4o-mini and its Qwen2-VL predecessor as the top open-source vanilla LMM on the MedXpertQA expert-level medical benchmark.

At a Glance

Where Qwen2.5-VL 72B leads

• Top open-source vanilla LMM score on MedXpertQA, ahead of GPT-4o-mini and Qwen2-VL-72B
• Fully open weights, enabling self-hosted deployment and fine-tuning

Where it lags

• Trails flagship proprietary models (GPT-4o) and dedicated reasoning models on complex tasks

Best for: teams needing a strong open-weight multimodal model with full deployment control.

What Qwen2.5-VL 72B Is

Qwen2.5-VL 72B is Alibaba's flagship open-weight multimodal model in the Qwen2.5-VL family, improving on Qwen2-VL with better visual understanding, document parsing, and agentic capabilities.

Specifications