AI · Official
Agent Lightning v1.0 Released: A Lightweight 3,500-Line Framework for Training AI Agents in Real Settings
Microsoft Research blog ·
At a glance
Microsoft Research Asia has introduced a new way to train AI agents directly in the real working environments where they operate. The open-source tool Agent Lightning v1.0, which implements this approach, is lightweight at about 3,500 lines of code. Its key feature is that it can train an agent that is already in use without rebuilding it.
Key points
- Previous methods required rebuilding the agent inside the training tool, which was costly and could make it differ from the agent actually deployed.
- The new tool trains agents without changing their existing code, routing them through a device that mediates access to the model.
- It supports Kubernetes (a management system that automatically splits and runs work across many computers) by default, so it can run on private servers or in the cloud.
- In a coding agent experiment, training on only about 6,000 examples raised one model's SWE-bench Verified accuracy from 41.8% to 56.4%, a gain of 14.6 percentage points.
Why it matters
As AI agents are increasingly used for real work, training methods that make them smarter are becoming more important. This tool aims to make that process cheaper and simpler.
Original source
Agent Lightning v1.0: A 3,500-Line Lightweight Agentic RL Framework for Training Agents with Real Harnesses
Microsoft Research blog · Oct 8, 01:00 KST
Read on Microsoft Research blog ↗AI summarized the original article. Check the original for accuracy. Copyright belongs to Microsoft Research blog.
More from this field
- AWS Marketplace Lets AI Agents Buy Software on Their Own
AWS has updated its AWS Marketplace so AI agents can buy software, APIs, data, and content on their own, not just find them. Buyers pay in dollars, and charges are added to their AWS bill.
디지털투데이 · Nov 21, 07:37 KST - Amazon SageMaker HyperPod Global Inference Gateway Routes AI Requests Across Regions
Amazon SageMaker HyperPod now has a Global Inference Gateway that sends each request to the best cluster across regions through one endpoint. Before this, clients had to choose a cluster themselves, and a busy cluster could leave requests waiting while idle GPUs sat elsewhere.
Amazon AWS Machine Learning blog · Nov 21, 07:35 KST - Google Drops Search Button and Adds AI Sidebar to Its Search Homepage
Google has officially rolled out a redesigned desktop search homepage after a successful test. The Google Search button next to the search bar is gone, and a row of AI buttons now sits under the bar. Logged-in users also get a sidebar showing their recent AI Mode conversations.
Ars Technica · Nov 21, 06:58 KST - Meta Rethinks 2030 Net-Zero Goal as AI Data Center Emissions Hit Record High
Meta's 2026 sustainability report said the company is re-evaluating its 2030 net-zero goal. Its emissions rose about 40 percent from 2024 to 2025, the largest jump among the big US tech companies.
Ars Technica · Nov 21, 06:47 KST