List your company

Multilingual Reward Modeling on m-RewardBench Avg. 23 langs

91.37AccuracyDIBJudge-Qwen3-8B
65.99472.58279.1785.758Mar 11, 2026

Evaluation results

MethodLinks
DIBJudge-Qwen3-8BBackbone=Qwen3-8B2026.0391.37
DIBJudge-Qwen3-4BBackbone=Qwen3-4B2026.0389.84
Nemotron-Multi-49BModel Category=Multilingual Open Reward Models, Parameters=49B2026.0388.83
mR3-Qwen3-8BModel Category=Multilingual Open Reward Models, Backbone=Qwen3-8B2026.0388.58
Gemini-2.5-FlashModel Category=Proprietary Models2026.0388.06
mR3-Qwen3-4BModel Category=Multilingual Open Reward Models, Backbone=Qwen3-4B2026.0387.21
Qwen3-8BModel Category=General Open Models, Parameters=8B2026.0386.12
GPT-4oModel Category=Proprietary Models2026.0385.75
Qwen3-4BModel Category=General Open Models, Parameters=4B2026.0385.06
Think-as-Locals 7BModel Category=Multilingual Open Reward Models, Parameters=7B2026.0384.51
M-PROMETHEUS 7BModel Category=Multilingual Open Reward Models, Parameters=7B2026.0378.03
Qwen2.5-7B-InstructModel Category=General Open Models, Parameters=7B2026.0377.89
M-PROMETHEUS 3BModel Category=Multilingual Open Reward Models, Parameters=3B2026.0368.45
Qwen2.5-3B-InstructModel Category=General Open Models, Parameters=3B2026.0366.97