Alibaba Qwen Releases QVQ-Max: Visual Reasoning Model with Proofs
Alibaba/Qwen
Alibaba Qwen officially releases QVQ-Max, its first version of a visual reasoning model. The model is capable of not only recognizing images and videos but also analyzing them, conducting reasoning, and offering solutions. In the MathVision benchmark, QVQ-Max demonstrates consistent accuracy improvement as the length of the thought process increases.
Alibaba's Qwen has introduced QVQ-Max, its first version of a visual reasoning model. The model can analyze and reason based on images and videos, solving tasks ranging from mathematical to everyday problems. In the MathVision benchmark, which includes diverse multimodal math tasks, QVQ-Max shows improved accuracy as the maximum thought chain length increases. Key capabilities of the model include detailed observation, deep reasoning, and flexible application. Use cases include assistance in work, learning, and daily life. Future plans involve improving recognition accuracy, multi-step task execution, and interaction with other modalities.
Source: Alibaba Qwen —
original