context.vn
Menu

mmlu Redux

11 models evaluated

#ModelProviderTypeScore
1Claude Opus 4.5AnthropicProprietary96.6%
2Qwen3.7 MaxAlibabaProprietary95%
3Qwen3.5 397BAlibabaOpen weights94.9%
4Qwen3.7 PlusAlibaba · Closed94.5%
5Qwen3.6 PlusAlibaba · Closed94.5%
6Qwen3.6-27BAlibaba · Open weight93.5%
9Ternary Bonsai 2 27BPrism ML · Open weight89.1%
10Mellum2-12B-A2.5B-ThinkingJetBrains · Open weight86.2%
11MiniCPM5-2BOpenBMB · Open weight84.7%
12Mellum2-12B-A2.5B-InstructJetBrains · Open weight78.1%
13MiniCPM5-1BOpenBMB · Open weight70.1%