Skip to content

Benchmark

SWE Refactor Bench

Coding benchmarks compare models on programming tasks. Scores are percentages. Higher is better.

Models measured
1
Scores recorded
2
Sources
2
IQ weight
No weight published

Score distribution

23%
Mean
23%
Median
0%
Std dev
23%
10th percentile
23%
90th percentile
23%
Min

Leaderboard

#ModelScore
1MiniMax M3.1 Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim23%
2MiniMax M3.1 FlashMiniMax M3.1 Flash official-weights model-card run · reasoning effort, sampling, and run count not stated for this benchmark · Claude Code 2.1.220 · 200K context · otherwise official setup· Reported by MiniMax· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.23%

Data credits: BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links. · Data by LLM Stats