Pith. sign in

REVIEW 13 cited by

R2E-Gym: Procedural Environments and Hybrid Verifiers for Scaling Open-Weights SWE Agents

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2504.07164 v1 pith:T3ZH7MLM submitted 2025-04-09 cs.SE cs.CLcs.LG

classification cs.SEcs.CLcs.LG
keywords environmentsmodelsscalingverifierscurationscalabletest-timeagentgym
verification ladder T0 review T1 audit T2 compute T3 formal

Signed reviews

No signed human review yet.

0 comments
read the original abstract

Improving open-source models on real-world SWE tasks (solving GITHUB issues) faces two key challenges: 1) scalable curation of execution environments to train these models, and, 2) optimal scaling of test-time compute. We introduce AgentGym, the largest procedurally-curated executable gym environment for training real-world SWE-agents, consisting of more than 8.7K tasks. AgentGym is powered by two main contributions: 1) SYNGEN: a synthetic data curation recipe that enables scalable curation of executable environments using test-generation and back-translation directly from commits, thereby reducing reliance on human-written issues or unit tests. We show that this enables more scalable training leading to pass@1 performance of 34.4% on SWE-Bench Verified benchmark with our 32B model. 2) Hybrid Test-time Scaling: we provide an in-depth analysis of two test-time scaling axes; execution-based and execution-free verifiers, demonstrating that they exhibit complementary strengths and limitations. Test-based verifiers suffer from low distinguishability, while execution-free verifiers are biased and often rely on stylistic features. Surprisingly, we find that while each approach individually saturates around 42-43%, significantly higher gains can be obtained by leveraging their complementary strengths. Overall, our approach achieves 51% on the SWE-Bench Verified benchmark, reflecting a new state-of-the-art for open-weight SWE-agents and for the first time showing competitive performance with proprietary models such as o1, o1-preview and sonnet-3.5-v2 (with tools). We will open-source our environments, models, and agent trajectories.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 13 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Scrouting: Cost-Aware Routing of Coding Agents by Scouting the Repository First

    cs.SE 2026-08 conditional novelty 6.0 of 10

    A 7B searcher's sandbox-verified handoff lets a cheap coding model match a frontier model's solve rate on SWE-bench Pro at roughly one fifth the cost, while the router that picks the fixer contributes nothing on this ...

  2. Fail-Fast, Restart-Smart: Early Failure Prediction and Restart for SWE Agentic Tasks

    cs.SE 2026-08 conditional novelty 6.0 of 10

    A prefix-only 0.6B monitor can forecast SWE-agent failure and, paired with an edit-overlay restart, reclaim tokens and improve resolution.

  3. RSIBench-Data: Benchmarking Data-Centric Research for Recursive Self-Improvement

    cs.SE 2026-07 conditional novelty 6.0 of 10

    A controlled benchmark shows LLM agents can sometimes discover better training-data strategies through feedback, but their improvements are fragile and usually not sustained.

  4. NexForge: Scaling Agent Capabilities through Requirement-Driven Task Synthesis for LLMs

    cs.SE 2026-07 conditional novelty 6.0 of 10

    Requirement-driven synthesis generates executable agent training tasks and trajectories from a capability description, improving a 35B model from 22.5% to 52.0% on Terminal-Bench 2.0 and from 813 to 1338 GDPval Elo.

  5. SWE-Review: Closing the Loop on Issue Resolution with Agentic Code Review

    cs.SE 2026-07 conditional novelty 6.0 of 10

    An agentic code reviewer that explores repositories to judge and diagnose AI-generated pull requests improves resolve rates from 27.5% to 56.9% and outperforms single-turn review baselines.

  6. BeyondSWE: Can Current Code Agent Survive Beyond Single-Repo Bug Fixing?

    cs.CL 2026-03 accept novelty 6.0 of 10

    A 500-task benchmark shows AI code agents plateau around 56% accuracy on broader software tasks, with search access providing limited and inconsistent gains.

  7. SERA: Soft-Verified Efficient Repository Agents

    cs.CL 2026-01 conditional novelty 6.0 of 10

    SERA trains competitive coding agents via supervised finetuning on synthetic trajectories verified only by patch-overlap recall, cutting data-generation cost by roughly two orders of magnitude.

  8. Toward Training Superintelligent Software Agents through Self-Play SWE-RL

    cs.SE 2025-12 unverdicted novelty 6.0 of 10

    Self-play RL on bug injection and repair in sandboxed repositories yields +10.4 and +7.8 point gains on SWE-bench Verified and Pro while outperforming human-data baselines.

  9. Satori-SWE: Evolutionary Test-Time Scaling for Sample-Efficient Software Engineering

    cs.CL 2025-05 conditional novelty 6.0 of 10

    A 32B model trained to self-evolve patch candidates through RL reaches 41.6% on SWE-Bench Verified with 50 samples, comparable to a 70B model using 500 samples.

  10. Process Reward Informed Tree Rollout for Effective Multi-Turn RL

    cs.CL 2026-07 conditional novelty 5.0 of 10

    Using process feedback to guide which partial trajectories to branch on, PATR replaces uniform independent rollouts with adaptive tree rollouts for multi-turn agent RL, reporting +5.0 points on SWE-Bench and +9.3 on F...

  11. TRACE: Turn-level Reward Assignment via Credit Estimation for Long-Horizon Agents

    cs.LG 2026-07 conditional novelty 5.0 of 10

    Turn-level rewards from TD changes in a frozen reference model's gold-answer log-probability improve long-horizon search-agent RL on closed- and open-web benchmarks.

  12. LiteResearcher: A Scalable Agentic RL Training Framework for Deep Research Agent

    cs.AI 2026-04 unverdicted novelty 5.0 of 10

    LiteResearcher uses a lite virtual world to make agentic RL training scalable and stable, enabling a 4B model to achieve 71.3% on GAIA and 78.0% on Xbench, outperforming larger open-source and commercial systems.

  13. Self-Evolving Coding Agents

    cs.SE 2026-08 conditional novelty 4.0 of 10

    A systematic survey that defines self-evolving coding agents, organizes systems into a five-category object-centered taxonomy, and analyzes when and on what evidence these agents evolve.

Pith tools