That DeepSeek R1 Distill example was surprising. Beating GPT-4o on reasoning with a 7B model isn’t small news.

→ View original post on X — @godofprompt