Alibaba Qwen2.5-VL-32B: Smaller Vision Model, Higher Scores

Alibaba released a 32-billion-parameter vision-language model under Apache 2.0, saying it beat the larger 72B Qwen model on math and image-understanding benchmarks.