DeepSeek R1 Distill Llama 70B

An open-weight 70B reasoning model produced by distilling DeepSeek-R1's chain-of-thought behavior into Meta's Llama-3.3-70B-Instruct backbone. It targets math, coding, and logic tasks with strong reasoning quality at a fraction of the compute cost of the full 671B DeepSeek-R1. Source: supabase.