Morning Edition · Tuesday, October 13, 2026Published at 2:13 AM EDT · New York

Alibaba's QwenGyre cuts agent reinforcement-learning wall-clock time by up to 1.85x by moving GPUs between rollout and training mid-run

On the 2.4-trillion-parameter Qwen 3.8, with rollouts of up to 700,000 tokens, Qwen reports a 6-point gain on NL2RepoBench, from 52.5% to 58.5%, in 48 training steps.

Alibaba's Qwen team released QwenGyre, a reinforcement-learning framework for training agents on tasks that run for hours and generate hundreds of thousands of tokens. AI ML Big Data described the problem it targets. In agent reinforcement…