deepseek-ai/DeepSeek-R1-Distill-Llama-70B-free

The first reasoning model from DeepSeek. Outperforms OpenAI GPT-4-o1 on multiple benchmarks.