Kimi K2 Base

Kimi K2 Base is a cutting-edge language model built on the mixture-of-experts (MoE) architecture with 32 billion activated parameters and 1 trillion total parameters. Trained on 15.5 trillion tokens using the MuonClip optimizer, this is a foundational model prior to instruction tuning. It demonstrates strong performance on knowledge, reasoning, and coding benchmarks while being optimized for agentic capabilities.