Kimi Linear 48B A3B Instruct

KimiKimiChina

Kimi Linear 48B A3B Instruct is a large language model optimized for efficiency, likely utilizing a Mixture-of-Experts (MoE) architecture with 48 billion total parameters and 3 billion active parameters. The 'Linear' designation suggests it may employ linear attention mechanisms for enhanced performance on long sequences. As an instruction-tuned model, it is designed for strong instruction following and conversational capabilities.