Open Source · MarkTechPost ·

Kimi AI and kvcache-ai Open Sources ‘AgentENV’: A Distributed System that Powers Agentic Reinforcement Learning (RL) Training for Kimi K3

Kimi AI and kvcache-ai Open Sources ‘AgentENV’: A Distributed System that Powers Agentic Reinforcement Learning (RL) Training for Kimi K3

Moonshot AI’s Kimi team and kvcache-ai released AgentENV under the MIT license. The distributed system runs agent sandboxes in Firecracker microVMs, supporting millisecond snapshots, resumes, 16-way forking, and an E2B-compatible API for reinforcement learning training.

Read the full story at MarkTechPost →