Hacker News new | ask | show | jobs
by neotrope 1142 days ago
It’s far better at producing workable code, logical reasoning, keeping a train of thought, etc. Their benchmark breaks it down.

https://openai.com/research/gpt-4