mlsanityData to 14 Oct 2026

QwenGyre: An Elastic Reinforcement Learning Framework for Training xLong-Horizon Agents

arXiv 2609.33848 ↗First posted 27 Sep 2026cs.LG cs.DCcited by 0

On the map: LLM agents and tool use → Reinforcement learning for LLM agents

also in 1 more
Abstract (arXiv)

Large language model (LLM) agents increasingly undertake extreme-long (xlong) horizon tasks, where a single execution can span hours, hundreds of model--environment interactions, and nearly 1M tokens per rollout. Applying online reinforcement learning (RL) to such executions poses two fundamental challenges: (1) severe execution variance and prolonged rollout delays cause massive GPU idling; and (2) complex non-linear branching generates massive trajectory redundancy, crippling training efficiency. To address these, we presents QwenGyre, an end-to-end framework for xlong-horizon online RL. QwenGyre elastically reallocates GPUs between rollout and training without interrupting live executions, while its trajectory processor reconstructs branching histories, scores partial progress, and deduplicates redundant paths to bound training costs. Scaled to our flagship model, Qwen~3.8 2.4T, with 700K tokens per rollout, QwenGyre yields a 6.0% absolute gain on NL2RepoBench (52.5% → 58.5%) in 48 steps. Across our evaluations on diverse domains of training datasets, QwenGyre delivers up to 1.85× and 1.78× speedups over Colocate and Async, respectively.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Reinforcement learning for LLM agents, one of 16 directions in LLM agents and tool use. Builds on 43 of its 57 references that are in the map. First posted on arXiv 27 Sep 2026.

Builds on

43 of 57 references.

Closest ancestors · 13

Same direction, newest first

  1. 2026SPO++: Stream-Aligned Policy Optimization for Asynchronous Agentic RL25 Aug
  2. 2026IAPO: Influence-Aware Policy Optimization for Credit Assignment in Multi-Turn Service Agents25 Aug
  3. 2026MileGPO: Milestone Inference with Local Evidence for Graph-Based Policy Optimization of Long-Horizon LLM Agents20 Aug
  4. 2026TideRL: Boosting Agentic RL Goodput with Readiness-Aware Scheduling11 Aug
  5. 2026Branching Policy Optimization: Sandbox-Native Language Agent Reinforcement Learning15 Jul
  6. 2026Single-Rollout Asynchronous Optimization for Agentic Reinforcement Learning8 Jul
  7. 2026Libra: Efficient Resource Management for Agentic RL Post-Training2 Jun
  8. 2026Heddle: A Distributed Orchestration System for Agentic RL Rollout30 Mar
  9. 2025SkyRL-Agent: Efficient RL Training for Multi-turn LLM Agent20 Nov
  10. 2025AgentRL: Scaling Agentic Reinforcement Learning with a Multi-Turn, Multi-Task Framework5 Oct
  11. 2025VerlTool: Towards Holistic Agentic Reinforcement Learning with Tool Use1 Sep
  12. 2025Agent Lightning: Train ANY AI Agents with Reinforcement Learning5 Aug
1 more
  1. 2025RAGEN: Understanding Self-Evolution in LLM Agents via Multi-Turn Reinforcement Learning24 Apr

Other references · 30

  1. 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,695
  2. 2023SWE-bench: Can Language Models Resolve Real-World GitHub Issues?cited by 3,373
  3. 2024HybridFlow: A Flexible and Efficient RLHF Frameworkcited by 2,057
  4. 2019Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelismcited by 1,868
  5. 2024SWE-agent: Agent-Computer Interfaces Enable Automated Software Engineeringcited by 1,737
  6. 2026Terminal-Bench: Benchmarking Agents on Hard, Realistic Tasks in Command Line Interfacescited by 585
  7. 2025SWE-Bench Pro: Can AI Agents Solve Long-Horizon Software Engineering Tasks?cited by 279
  8. 2025AReaL: A Large-Scale Asynchronous Reinforcement Learning System for Language Reasoningcited by 250
  9. 2025Reinforcement Learning Optimization for Large-Scale Learning: An Efficient and User-Friendly Scaling Librarycited by 101
  10. 2025NL2Repo-Bench: Towards Long-Horizon Repository Generation Evaluation of Coding Agentscited by 65
  11. 2025StreamRL: Scalable, Heterogeneous, and Elastic RL for LLMs with Disaggregated Stream Generationcited by 61
  12. 2026DeepSWE: Measuring Frontier Coding Agents on Original, Long-Horizon Engineering Taskscited by 49
18 more
  1. 2025AsyncFlow: An Asynchronous Streaming RL Framework for Efficient LLM Post-Trainingcited by 41
  2. 2024ReaL: Efficient RLHF Training of Large Language Models with Parameter Reallocationcited by 37
  3. 2026SWE-Master: Unleashing the Potential of Software Engineering Agents via Post-Trainingcited by 22
  4. 2026DORA: A Scalable Asynchronous Reinforcement Learning System for Language Model Trainingcited by 14
  5. 2026Polar: Agentic RL on Any Harness at Scalecited by 14
  6. 2026StaleFlow: Staleness-Aware Data Management for Mitigating Data Skewness in Fully Disaggregated RL Post-Trainingcited by 11
  7. 2026MegaFlow: Large-Scale Distributed Orchestration System for the Agentic Eracited by 9
  8. 2026ROSE: Rollout On Serving GPUs via Cooperative Elasticity for Agentic RLcited by 8
  9. 2026Agent Lightning v1.0: Towards Harnessed Agentic RLcited by 7
  10. 2025Tree Training: Accelerating Agentic LLMs Training via Shared Prefix Reusecited by 5
  11. 2026SWE Refactor Bench: Can Coding Agents Complete a Long-Horizon, Whole-Repository Stack Migration?cited by 5
  12. 2026RolloutPipe: Overlapping Pipelined Rollout and Training in Disaggregated On-Policy LLM Reinforcement Learningcited by 4
  13. 2026LEGO-RL: Harness-Native Reinforcement Learning for Coding Agentscited by 4
  14. 2026AReaL-DTE: Sparse Policy-Weight Transfer for Online Agentic Reinforcement Learningcited by 3
  15. 2026DynaResize: Runtime GPU Reallocation for Disaggregated LLM Post-Trainingcited by 2
  16. 2026ClawGym II: Exploring Black-Box RL on Agent Harnesscited by 2
  17. 2026TailSieve: Partial-Rollout-Guided Tail Routing for LLM Rolloutscited by 2
  18. 2026psRL: Efficient Training for Agentic AI via Training-Time Prefix Sharingcited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Miles v0.1: Production-Level Post-Training
  2. 2026Demystifying Reinforcement Learning for Long-Horizon Tool-Using Agents: A Comprehensive Recipe
  3. 2025QeRL: Beyond Efficiency -- Quantization-enhanced Reinforcement Learning for LLMs
  4. 2025Laminar: A Scalable Asynchronous RL Post-Training Framework
  5. 2025The Art of Scaling Reinforcement Learning Compute for LLMs
  6. 2025EARL: Efficient Agentic Reinforcement Learning Systems for Large Language Models
  7. 2026From Outcomes to Actions: Leveraging Hindsight for Long-Horizon Language Agent Training
  8. 2026ThunderSyncRL: Lossless Acceleration of Agentic Reinforcement Learning