Entity · product

R2E-Gym

productactiveprovisionalr2e-gym-3bf8ed19·1 events·first seen 6d ago

Aliases: R2E-Gym

Co-occurring entities

SWE-Smith SWE-Bench Lite Qwen2.5-Coder-32B-Instruct SWE-Bench Verified Qwen3-4B LiveCodeBench BFCL Function-Aware Fill-in-the-Middle as Mid-Training for Coding Agent Foundation Models TAU-bench

More like this (12)

NatureGym MobileGym EvoPolicyGym CyberGym DSGym Gym Retro R2R-CE Safety Gym MobileGym-Bench R2R OpenAI Gym R2 Indicator

Recent events (1)

6arXiv · cs.CL·6d ago·source ↗

Function-Aware Fill-in-the-Middle Mid-Training Improves Coding Agent Foundation Models

Researchers propose a self-supervised mid-training objective called function-aware fill-in-the-middle (FIM) that exploits the structural isomorphism between a coding agent's action-observation-continuation loop and function call sites in ordinary code. Applied to Qwen2.5-Coder-Instruct (7B/14B) and Qwen3-8B on a 2.6B-token GitHub corpus, the method yields +2.8 to +5.4 point gains on SWE-Bench-Verified and SWE-Bench-Lite across multiple post-training pipelines. Notably, the technique also mitigates capability erosion on non-agent coding and tool-use benchmarks, suggesting the function-call inductive bias generalizes beyond the training domain.

Frontier Model Releases Evaluation and Benchmarking SWE-Smith SWE-Bench Lite Qwen2.5-Coder-32B-Instruct +8 more