SmellBench Refactoring (OpenHands) - Structural Soundness: leaderboard

Metric: Structural soundness score (0-100): correctness and rationality of the refactored structure; an LLM judge (model not named) scores the refactoring against the smell analysis on a 0-10 rubric, normalized to 0-100; 294 injected-smell refactoring cases (7 smell types, 3 difficulty levels, guided and targeted instructions) in 7 real Python repositories, run once in the Harbor Docker framework with a 20-minute budget, OpenHands agent harness; higher is better. Source: arxiv.org. Saturation forecast: Around December 2026. 6 models tracked.

Top models

#ModelScore
1Claude Sonnet 4.561.39
2GPT-5 Mini54.18
3DeepSeek V3.252.03
4Qwen 3 Coder 480B A35B Instruct50.84
5Qwen 3 Coder 30B A3B Instruct41.1
6Gemini 2.5 Flash35.92

Interactive version: theaggregate.ai/benchmark?slug=smellbench-refactoring-openhands-structural-soundness · How It Works · Data refreshed daily, snapshot 2026-09-29.