月之暗面发布 Kimi K3 模型权重与技术报告
Releasing the model weights and technical report of Kimi K3.
AI 导读
月之暗面发布 Kimi K3 的模型权重与技术报告。Kimi K3 是其目前能力最强的模型，采用 2.8T MoE 架构，具备原生视觉理解和 1M token 上下文窗口，官方称新架构带来单位算力 2.5 倍的智能提升。随 Kimi K3 一同开放的还有高性能 attention kernel、MoE 通信库以及大规模运行智能体环境的基础设施。
正文
当前显示原文，中文全文译文尚未就绪。
Releasing the model weights and technical report of Kimi K3.
Kimi K3 is our most capable model: a 2.8T MoE model with native visual understanding and a 1M-token context window.
New model architecture: 2.5x the intelligence per unit of compute, not just more params.
Alongside Kimi K3, we're opening up more of the stack behind it — high-performance attention kernels, MoE communication library, and infrastructure for running agent environments at scale.