List your company

ResearchFormerly WizWand

Multilingual Reward Modeling

Benchmarks

DatasetSOTA methodMetricTrendResultsLast updated

MM-Eval Avg. 18 lang

DIBJudge-Qwen3-8B

87.53Accuracy

14

Mar 12, 2026

m-RewardBench Avg. 23 langs

DIBJudge-Qwen3-8B

91.37Accuracy

14

Mar 12, 2026