# PipelineRL: Faster On-policy Reinforcement Learning for Long Sequence Generation

arXiv: https://arxiv.org/abs/2509.19128 · HTML: https://mlsanity.com/paper/2509.19128/

## Key facts

- Cited by 32 papers in the map (data through 14 Oct 2026).
- Placed in Reinforcement learning for reasoning, one of 17 directions in LLM reasoning.
- Builds on 20 of its 30 references that are in the map.
- First posted on arXiv 23 Sep 2025.

## Placed in

- LLM reasoning / [Reinforcement learning for reasoning](https://mlsanity.com/lr/reinforcement-learning-for-reasoning/)

## Builds on (most cited first, 20 of 20)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [Measuring Mathematical Problem Solving With the MATH Dataset](https://mlsanity.com/paper/2103.03874/) (2021)
- [Qwen2.5 Technical Report](https://mlsanity.com/paper/2412.15115/) (2024)
- [HybridFlow: A Flexible and Efficient RLHF Framework](https://mlsanity.com/paper/2409.19256/) (2024)
- [Dota 2 with Large Scale Deep Reinforcement Learning](https://mlsanity.com/paper/1912.06680/) (2019)
- [IMPALA: Scalable Distributed Deep-RL with Importance Weighted Actor-Learner Architectures](https://mlsanity.com/paper/1802.01561/) (2018)
- [SimpleRL-Zoo: Investigating and Taming Zero Reinforcement Learning for Open Base Models in the Wild](https://mlsanity.com/paper/2503.18892/) (2025)
- [Open-Reasoner-Zero: An Open Source Approach to Scaling Up Reinforcement Learning on the Base Model](https://mlsanity.com/paper/2503.24290/) (2025)
- [Safe and Efficient Off-Policy Reinforcement Learning](https://mlsanity.com/paper/1606.02647/) (2016)
- [SWE-RL: Advancing LLM Reasoning via Reinforcement Learning on Open Software Evolution](https://mlsanity.com/paper/2502.18449/) (2025)
- [Asynchronous RLHF: Faster and More Efficient Off-Policy RL for Language Models](https://mlsanity.com/paper/2410.18252/) (2024)
- [Tapered Off-Policy REINFORCE: Stable and efficient reinforcement learning for LLMs](https://mlsanity.com/paper/2503.14286/) (2025)
- [P3O: Policy-on Policy-off Policy Optimization](https://mlsanity.com/paper/1905.01756/) (2019)
- [Importance Resampling for Off-policy Prediction](https://mlsanity.com/paper/1906.04328/) (2019)

## Built on by (most cited first, 30 of 32)

- [POPE: Learning to Reason on Hard Problems via Privileged On-Policy Exploration](https://mlsanity.com/paper/2601.18779/) (2026)
- [Nemotron 3 Super: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoning](https://mlsanity.com/paper/2604.12374/) (2026)
- [Seer: Online Context Learning for Fast Synchronous LLM Reinforcement Learning](https://mlsanity.com/paper/2511.14617/) (2025)
- [Composer 2 Technical Report](https://mlsanity.com/paper/2603.24477/) (2026)
- [WebGym: Scaling Training Environments for Visual Web Agents with Realistic Tasks](https://mlsanity.com/paper/2601.02439/) (2026)
- [InT: Self-Proposed Interventions Enable Credit Assignment in LLM Reasoning](https://mlsanity.com/paper/2601.14209/) (2026)
- [Stable Asynchrony: Variance-Controlled Off-Policy RL for LLMs](https://mlsanity.com/paper/2602.17616/) (2026)
- [Bolmo: Byteifying the Next Generation of Language Models](https://mlsanity.com/paper/2512.15586/) (2025)
- [Accelerating RL Post-Training Rollouts via System-Integrated Speculative Decoding](https://mlsanity.com/paper/2604.26779/) (2026)
- [Aletheia: What Makes RLVR For Code Verifiers Tick?](https://mlsanity.com/paper/2601.12186/) (2026)
- [EMA Policy Gradient: Taming Reinforcement Learning for LLMs with EMA Anchor and Top-k KL](https://mlsanity.com/paper/2602.04417/) (2026)
- [Sparrow: Sparse Rollout for Stable and Efficient Long-context RL of Large Language Models](https://mlsanity.com/paper/2606.08446/) (2026)
- [Not All LLM Reasoning is Visible in the Chain-of-Thought](https://mlsanity.com/paper/2607.22925/) (2026)
- [GrandCode: Achieving Grandmaster Level in Competitive Programming via Agentic Reinforcement Learning](https://mlsanity.com/paper/2604.02721/) (2026)
- [ZAYA1-8B Technical Report](https://mlsanity.com/paper/2605.05365/) (2026)
- [FrogNano: Training a 4B Coding Agent via Online Task Synthesis](https://mlsanity.com/paper/2609.07925/) (2026)
- [QED-Nano: Teaching a Tiny Model to Prove Hard Theorems](https://mlsanity.com/paper/2604.04898/) (2026)
- [Score Centering Stabilizes Off-policy Reinforcement Learning](https://mlsanity.com/paper/2609.20807/) (2026)
- [Staleness-Learning Rate Scaling Laws for Asynchronous RLHF](https://mlsanity.com/paper/2607.01083/) (2026)
- [SpecRoll: Fast-Slow Verifier-Feedback Adaptation for Speculative Reinforcement Learning Rollouts](https://mlsanity.com/paper/2608.04962/) (2026)
- [When Do Larger Batches Help Scale LLM Reinforcement Learning?](https://mlsanity.com/paper/2608.29296/) (2026)
- [CLaaS: Continual learning as a service for sample efficient online learning](https://mlsanity.com/paper/2606.05559/) (2026)
- [Learning to Solve Hard Problems in RL for LLMs by Never Giving Up](https://mlsanity.com/paper/2609.13443/) (2026)
- [KV-streams for Efficient Compaction in Agentic Reinforcement Learning](https://mlsanity.com/paper/2609.35750/) (2026)
- [Apriel-Reasoner: RL Post-Training for General-Purpose and Efficient Reasoning](https://mlsanity.com/paper/2604.02007/) (2026)
- [Deconstructing Off-Policy Ratios: Entropy-Normalized Trust Regions for Asynchronous Reinforcement Learning](https://mlsanity.com/paper/2607.22186/) (2026)
- [BRACE: Anchored Bellman-Residual Correction for Stale Critics in Asynchronous RL](https://mlsanity.com/paper/2609.09783/) (2026)
- [An Open Recipe for IMO Gold: Training Nemotron for Olympiad Mathematics](https://mlsanity.com/paper/2609.10712/) (2026)
- [ABC: Advantage-Based Control Variates for Reinforcement Learning with Verifiable Rewards](https://mlsanity.com/paper/2609.36058/) (2026)
- [ThunderSyncRL: Lossless Acceleration of Agentic Reinforcement Learning](https://mlsanity.com/paper/2610.05935/) (2026)

## Abstract (arXiv)

Reinforcement Learning (RL) is increasingly utilized to enhance the reasoning capabilities of Large Language Models (LLMs). However, effectively scaling these RL methods presents significant challenges, primarily due to the difficulty in maintaining high AI accelerator utilization without generating stale, off-policy data that harms common RL algorithms. This paper introduces PipelineRL, an approach designed to achieve a superior trade-off between hardware efficiency and data on-policyness for LLM training. PipelineRL employs concurrent asynchronous data generation and model training, distinguished by the novel in-flight weight updates. This mechanism allows the LLM generation engine to receive updated model weights with minimal interruption during the generation of token sequences, thereby maximizing both the accelerator utilization and the freshness of training data. Experiments conducted on long-form reasoning tasks using 128 H100 GPUs demonstrate that PipelineRL achieves approximately ~ 2x faster learning compared to conventional RL baselines while maintaining highly on-policy training data. A scalable and modular open-source implementation of PipelineRL is also released as a key contribution.