New chat
Kimi-K2
Kimi-K2
By
155 followers
Kimi K2-Instruct-0905 is a state-of-the-art mixture-of-experts (MoE) language model with 32 billion activated parameters and 1 trillion total parameters. Trained with the Muon optimizer, Kimi K2 achieves exceptional performance across frontier knowledge, reasoning, and coding tasks while being meticulously optimized for agentic capabilities.
Key Features:
- Large-Scale Training: Pre-trained a 1T parameter MoE model on 15.5T tokens with zero training instability.
- MuonClip Optimizer: We apply the Muon optimizer to an unprecedented scale, and develop novel optimization techniques to resolve instabilities while scaling up.
- Agentic Intelligence: Specifically designed for tool use, reasoning, and autonomous problem-solving.
Technical Specifications
File Support: Attachments not supported
Context window: 256k tokens
Powered by a server managed by @novitaai. Learn more