AutoRL: Multi-Agent System Automates RL Training, Lifts Qwen3 AIME from 46.7% to 66.7%
omarsar0 · x · 2026-11-17
A paper from Meituan, CUHK, ShanghaiTech and Fudan introduces AutoRL, a multi-agent system that supervises live RL runs like an on-call RL researcher: Perceptor agents watch training curves and evals, three Proposers diagnose issues, and an Orchestrator applies hot updates or checkpoint restarts. Qwen3 improves on AIME 2024 from 46.67% to 66.67% (+11.67 on AIME/HMMT 2025); LongCat-Flash-Lite agentic coding pass@1 rises from 52.29% to 65.13%, sustaining 14.7 days of unattended training with 11 runtime failure fixes.
More from coding & agent
- 9 Hermes Agent Plugins Worth a Weekend, Picked From 2,300+ — Teknium · 2026-11-17
- Ant Group's Ling team cuts agent eval batch time by up to 65% with smarter scheduling — dair_ai · 2026-11-17
- Cloud vs local coding agents debate: "you've got a jumbo, why drive a VW Beetle?" — intellectronica · 2026-11-17
- Cloudflare Artifacts repos now testable locally via wrangler dev, with fork and push support — dinasaur_404 · 2026-11-17
- Courses are becoming SaaS: packaging your method as agent skills buyers can run — VibeMarketer_ · 2026-11-17
- Keyfleet launches persistent multiplayer workspaces where personal AI agents share goals and review each other — seanwbren · 2026-11-17