LATIDIA · Investigación

Agent Lightning v1.0: A 3,500-Line Lightweight Agentic RL Framework for Training Agents with Real Harnesses

Capacitar a los agentes de IA con aprendizaje de refuerzo puede ser un desafío porque sus herramientas, contexto y toma de decisiones se gestionan mediante marcos complejos. Agent Lightning conecta a los agentes existentes con la capacitación de RL, lo que lo convierte en un

La noticia

At a glance Harnessed Agentic RL: Microsoft Research Asia introduces a training paradigm in which the same agent harness used in deployment participates directly in reinforcement learning, removing the need to reimplement the agent inside the training framework. Lightweight by design: Agent Lightning v1.0 delivers a complete agent RL control plane in roughly 3,500 lines of code. Native Kubernetes support: agents run as standard Kubernetes jobs on self-managed clusters, cloud Kubernetes, or local infrastructure, with no dependency on paid commercial sandbox services. Data-efficient training recipe: an end-to-end coding agent pipeline raised Qwen3.5-9B from 41.8% to 56.4% Pass@1 on SWE-bench Verified, a 14.6 percentage point gain, using only about 6,000 training