LoopCoder-v2
LoopCoder-v2: Only Loop Once for Efficient Test-Time Computation Scaling
Trains a family of 7B Parallel Loop Transformer coders with different loop counts from scratch on 18T tokens to study loop-count selection as a gain–cost trade-off between extra refinement and the positional mismatch introduced by cross-loop position offsets.
Inside the method
PLT: [ Stack ] × 2 with cross-loop position offsets + shared-KV gated sliding window
Simplified conceptual schematic. Consult the paper for the complete architecture.
- Recurrence family
- Full-stack recurrence
- Depth control
- Two loops best; three or more regress
- KV / state strategy
- Shared KV with gated sliding-window attention (PLT)
Reading note
The two-loop model raises SWE-bench Verified from 43.0 to 64.4 and Multi-SWE from 14.0 to 31.0 over the non-looped baseline. Loop 2 provides the main refinement while later loops oscillate.
Sources checked 2026-10-28. This catalog does not imply independent reproduction.
Related works
Cite this work
@misc{loopcoderv22026,
title = {LoopCoder-v2: Only Loop Once for Efficient Test-Time Computation Scaling},
author = {Jian Yang and Shawn Guo and Wei Zhang and Tianyu Zheng and Yaxin Du and Haau-Sing Li and Jiajun Wu and Yue Song and Yan Xing and Qingsong Cai and Zelong Huang and Chuan Hao and Ran Tao and Xianglong Liu and Wayne Xin Zhao and Mingjie Tang and Weifeng Lv and Ming Zhou and Bryan Dai},
year = {2026},
eprint = {2606.18023},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2606.18023}
}