This survey introduces the Generate-Filter-Control-Replay (GFCR) taxonomy to structure rollout pipelines for RL-based post-training of reasoning LLMs.
Toward Training Superintelligent Software Agents through Self-Play SWE-RL
7 Pith papers cite this work. Polarity classification is still indexing.
7
Pith papers citing it
abstract
While current software agents powered by large language models (LLMs) and agentic reinforcement learning (RL) can boost programmer productivity, their training data (e.g., GitHub issues and pull requests) and environments (e.g., pass-to-pass and fail-to-pass tests) heavily depend on human knowledge or curation, posing a fundamental barrier to superintelligence. In this paper, we present Self-play SWE-RL (SSR), a first step toward training paradigms for superintelligent software agents. Our approach takes minimal data assumptions, only requiring access to sandboxed repositories with source code and installed dependencies, with no need for human-labeled issues or tests. Grounded in these real-world codebases, a single LLM agent is trained via reinforcement learning in a self-play setting to iteratively inject and repair software bugs of increasing complexity, with each bug formally specified by a test patch rather than a natural language issue description. On the SWE-bench Verified and SWE-Bench Pro benchmarks, SSR achieves notable self-improvement (+10.4 and +7.8 points, respectively) and consistently outperforms the human-data baseline over the entire training trajectory, despite being evaluated on natural language issues absent from self-play. Our results, albeit early, suggest a path where agents autonomously gather extensive learning experiences from real-world software repositories, ultimately enabling superintelligent systems that exceed human capabilities in understanding how systems are constructed, solving novel challenges, and autonomously creating new software from scratch.
years
2026 7representative citing papers
A self-evolving training loop that generates its own spatial QA data with executable code and difficulty feedback lifts Qwen3-VL-4B/8B to 62.7/63.3 on VSI-Bench using an order of magnitude less data.
PROPEL amortizes solver evaluation with a trained activation probe to optimize task generators toward a target solve rate, raising the share of learnable tasks from ~10% to ~20% in coding and SWE experiments.
Spreadsheet-RL applies RL fine-tuning and a custom Gym environment to raise LLM agent Pass@1 scores on spreadsheet benchmarks from roughly 8-12% to 17-23%.
EvoEnv lets a single policy synthesize, validate, and use Python environments with durable solve-verify asymmetry to improve reasoning performance on Qwen3-4B-Thinking from 72.4 to 74.8 while fixed-data baselines decline.
Preping builds agent memory via proposer-guided synthetic practice and selective validation, matching offline/online methods at 2-3x lower deployment cost.
Neural Change Prediction generates mutation data to train bidirectional models linking code changes to behavioral effects for any executable program.
citing papers explorer
-
Generate, Filter, Control, Replay: A Comprehensive Survey of Rollout Strategies for LLM Reinforcement Learning
This survey introduces the Generate-Filter-Control-Replay (GFCR) taxonomy to structure rollout pipelines for RL-based post-training of reasoning LLMs.
-
Ouroboros-Spatial: Closing the Data-Model Loop for Spatial Reasoning
A self-evolving training loop that generates its own spatial QA data with executable code and difficulty feedback lifts Qwen3-VL-4B/8B to 62.7/63.3 on VSI-Bench using an order of magnitude less data.
-
Breaking the Solver Bottleneck: Training Task Generators at the Learnable Frontier
PROPEL amortizes solver evaluation with a trained activation probe to optimize task generators toward a target solve rate, raising the share of learnable tasks from ~10% to ~20% in coding and SWE experiments.
-
Spreadsheet-RL: Advancing Large Language Model Agents on Realistic Spreadsheet Tasks via Reinforcement Learning
Spreadsheet-RL applies RL fine-tuning and a custom Gym environment to raise LLM agent Pass@1 scores on spreadsheet benchmarks from roughly 8-12% to 17-23%.
-
Learning to Build the Environment: Self-Evolving Reasoning RL via Verifiable Environment Synthesis
EvoEnv lets a single policy synthesize, validate, and use Python environments with durable solve-verify asymmetry to improve reasoning performance on Qwen3-4B-Thinking from 72.4 to 74.8 while fixed-data baselines decline.
-
PREPING: Building Agent Memory without Tasks
Preping builds agent memory via proposer-guided synthetic practice and selective validation, matching offline/online methods at 2-3x lower deployment cost.
-
Neural Change Prediction: Relating Software Changes to Their Effects and Vice Versa
Neural Change Prediction generates mutation data to train bidirectional models linking code changes to behavioral effects for any executable program.