Hacker News new | past | comments | ask | show | jobs | submit login

based on a few initial tests GPT-4.5 is abysmal. I find the prose more sterile than previous models and far from having the spark of DeepSeek, and it utterly choked on / mangled some python code (~200 LoC and 120 LoC tests) that o3-mini-high and grok-3 do very well on.



Consider applying for YC's Summer 2025 batch! Applications are open till May 13

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: