SPROUT: Scaling Agentic Reinforcement Learning for Long-Horizon Software Engineering
cs.SE, cs.CL, cs.LG
Submitted: 2026-10-09
Updated: 2026-10-09
Comments: work on process
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Large Language Monkeys: Scaling Inference Compute with Repeated Sampling
- SWE-Bench Pro: Can AI Agents Solve Long-Horizon Software Engineering Tasks?
- Branching Policy Optimization: Sandbox-Native Language Agent Reinforcement Learning
- OpenCollab: A Multi-Agent Coding Framework with Programmable Collaboration and Controllable Runtime
- DeepSWE: Measuring Frontier Coding Agents on Original, Long-Horizon Engineering Tasks
- VinePPO: Refining Credit Assignment in RL Training of LLMs
- TVCACHE: A Stateful Tool-Value Cache for Post-Training LLM Agents
- AgentGit: A Version Control Framework for Reliable and Scalable LLM-Powered Multi-Agent Systems
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
- Kimi K3: Open Frontier Intelligence
- Fork, Explore, Commit: OS Primitives for Agentic Exploration
- Executable Code Actions Elicit Better LLM Agents
- Contrastive Branch Policy Optimization
- Toward Systems Foundations for Agentic Exploration
- ReAct: Synergizing Reasoning and Acting in Language Models
- Shepherd: Enabling Programmable Meta-Agents via Reversible Agentic Execution Traces
- Language Agent Tree Search Unifies Reasoning Acting and Planning in Language Models
Related papers
- Falsification-Based Verification of LLM-Generated Optimization Models: Sound Test Batteries and Their Detection Limits
- GitSkills: A Dataset of Agent Skills on GitHub
- SABER: Benchmarking Operational Safety of LLM Coding Agents in Stateful Project Workspaces
- PackMonitor: Enabling Zero Package Hallucinations Through Decoding-Time Monitoring
- IntentCoding: Amplifying User Intent in Code Generation
- Incentives and Outcomes in Bug Bounties