According to the announcement of the Qwen team, the Qwen 2.5-VL-32B-Instruct model is officially open-sourced and has a 32B parameter scale, showing excellent performance in tasks such as image understanding, mathematical reasoning and text generation. The model is further optimized through reinforcement learning, and the responses are more in line with human preferences. It surpasses the previously released 72B model in multi-modal evaluations such as MMMU and MathVista.