Comment by krackers

Comment by krackers 4 hours ago

2 replies

I can believe this, Deepseek V3.2 shows that you can get close to "gpt-5" performance with a gpt-4 level base model just with sufficient post-training.