TechFlame News – The Kimi.ai team, while open-sourcing the Kimi K3 model weights and technical report, has also gradually released multiple underlying infrastructure components, including MoonEP, a high-performance communication library for distributed MoE training; AgentENV, a distributed environment system for large-scale agent workflows (in collaboration with kvcache-ai); and FlashKDA, a high-performance kernel based on CUTLASS for Kimi Delta Attention. According to the team, these components can reduce communication and inference overhead during large-scale MoE and agent reinforcement learning training, and can be used as a plug-and-play backend for flash-linear-attention.