ResearchBenchmarksUnderstandability on Understandability Experiment Code strategyUnderstandability on Understandability Experiment Code strategyFollow7Significant Results CountClaude2.843.9256.08May 7, 2026Evaluation resultsMethodMethodLinksSignificant Results CountAdjusted R2Spearman Correlation SignificanceMajority FitClaude2026.0570.86——GPT-4o-m2026.0560.99——GPT-4o2026.0550.76——Llama2026.0551——Grok2026.0550.83——Majority Vote (MUM)2026.055———Mistral2026.0531——Qwen2026.0531——