Hacker News new | ask | show | jobs
by darknoon 305 days ago
It's ok, somewhere between a qwen 2.5 VL and the frontier models (o3 / opus 4) on visual reasoning