Post by merve on X
merve@mervenoyann
XQwen just dropped Qwen2.5-VL-32B-Instruct ☄️
> 32B vision LM > Qwen-2.5-VL-72B in math, reasoning, RAG
> better aligned for human preferences
> same architecture, available in @huggingface transformers
find model and demo on the next one ⤵️

186 likes8 repliesPosted Mar 24, 2025