DeepSeek-R1-0528

DeepSeek-R1-0528 is an updated open-weight reasoning LLM from DeepSeek, built on a Mixture-of-Experts backbone and optimized for math, code, tool use, and multi-step problem solving. It is a checkpoint upgrade over the original R1, with reported improvements in reasoning depth, hallucination reduction, and function-calling accuracy. Source: supabase.