DeepSeek open-sourced DeepSeek-R1, an LLM fine-tuned with support knowing (RL) to enhance reasoning ability. DeepSeek-R1 attains results on par with OpenAI’s o1 model on a number of benchmarks, consisting of MATH-500 and SWE-bench.

DeepSeek-R1 is based on DeepSeek-V3, wiki.asexuality.org a mixture of professionals (MoE) model just recently open-sourced by DeepSeek. This base design is fine-tuned utilizing Group Relative (GRPO), a reasoning-oriented variation of RL. The research study group also performed knowledge distillation from DeepSeek-R1 to open-source Qwen and Llama models and launched several variations of each