Merge-Bench (All Languages): leaderboard

Merge-Bench (2026) gives models real merge-conflict hunks from open-source repositories and compares the returned resolution with the one the developers committed. This board covers all 11 programming languages and counts exact text matches, in percent; higher is better.

Metric: Equivalent text (%): share of conflicts whose output matches the developer resolution exactly as a string, all 7,938 merge-conflict hunks from 1,439 repositories in 11 languages, the model returns the resolved snippet in Markdown, compared with the resolution the developers committed; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 6 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Gemini 2.5 Pro47.1#276
2Claude Opus 440.3#326
3O3 Pro39.2#209
4DeepSeek R1 052832#428
5Grok 427.7#319
6Qwen 3 235B A22B25.8#628

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=merge-bench-all-languages · How It Works · Data refreshed daily, snapshot 2026-11-21.