context.vn
Menu

long Bench V2

14 models evaluated

#ModelProviderTypeScore
1Qwen3.8 MaxAlibabaOpen weights66.3%
2Claude Opus 4.5AnthropicProprietary64.4%
3Qwen3.5 397BAlibabaOpen weights63.2%
4Qwen3.6 PlusAlibaba · Closed62%
5Nemotron 3 UltraNVIDIA · Open weight61.9%
6Kimi K2.5Moonshot AI · Open weight61%
7GLM-5Z.AI · Open weight60.8%
8Qwen3.5-27BAlibaba · Open weight60.6%
9Qwen3.5-122B-A10BAlibaba · Open weight60.2%
10Agents-A1InternScience · Open weight60.2%
11Qwen3.5-35B-A3BAlibaba · Open weight59%
12Agents-A1-4BInternScience · Open weight52.1%
15MiniCPM5-2BOpenBMB · Open weight43.7%
16LLaDA2.2-miniInclusionAI · Open weight35.0%