QwenGyre: elastic RL for ultra-long-horizon agents
Alibaba's Qwen team released QwenGyre, an online reinforcement-learning framework for training agents on tasks that run for hours and burn ~700k tokens per rollout.
Why it matters
It attacks the core bottleneck of long-horizon agent RL — GPUs sitting idle while waiting on hours-long rollouts — with elastic scheduling that cut training time ~44% on a 2.4T model in published experiments.
What you could build with it
An ML platform team could adopt QwenGyre's elastic rollout-training scheduler to train coding agents on full repository-level tasks overnight — turning GPU time that used to sit idle waiting for long agent runs into actual training steps.
Does it hold up?
Too early to judge: the headline numbers (44% faster, +6pt on NL2RepoBench) are the lab's own, the paper dates to Sept 27, and no public code repository could be found on GitHub — real-world validation awaits the open-source release.
Built with QwenGyre
- Alibaba Unveils QwenGyre to Speed Up Training of AI Agentsfourweekmba · Walkthrough of the two GPU-idling problems QwenGyre attacks, the 75.42h vs 134.55h training comparison on Qwen3.8-2.4T, and the X announcement from the Qwen team.
- Alibaba Qwen Introduces QwenGyre: Elastic RL for XLong-Horizon Agentsmarktechpost · Detailed numbers: 700K tokens per rollout, 1.21-1.85x speedups over colocate/async baselines, NL2RepoBench 52.5% to 58.5% in 48 steps; notes no public code release was linked.
- Alibaba's Qwen introduces QwenGyre for long-horizon agent trainingaiunderstanding · Independent read: the practical meaning is that agent quality is moving from the model alone to the harness and training loop; flags that access, license, and pricing are undisclosed.
Learn more
First spotted on article: source.
More AI agents
Agent ReachOpen-source CLI and SKILL.md kit that gives AI agents read/search access to 15+ internet platforms (Twitter, Reddit…agents · JEV 0.8AI Job Search — the job hunt that runs on your machineOpen-source Claude Code framework that turns job hunting into an agentic pipeline: scrape portals, score fit, tailor…agents · JEV 0.79PentAGI — fully autonomous AI agents for penetration testingPentAGI is a self-hosted, open-source system where a team of AI agents plans and executes authorized penetration tests…agents · JEV 0.77PPT Master: AI skill that generates native, editable PowerPoint decks from any documentOpen-source AI agent workflow that turns documents or topics into real, native PowerPoint decks: native shapes…agents · JEV 0.76Docker Agent: AI agent builder and runtime by Docker EngineeringA Docker CLI plugin for building, running and sharing AI agents with declarative YAML config, multi-agent…agents · JEV 0.75K-Dense-AI/scientific-agent-skills: turn any AI agent into an AI scientistK-Dense AI's open collection of 177 validated, lazily-loaded skills that turn any agent (Cursor, Claude Code, Codex…agents · JEV 0.74
Get the week's best AI launches, plus 3 ideas worth building
One email every Saturday. Ranked by traction, not hype. Free.