context.vn

hmmt Feb2026

20 models evaluated

#ModelProviderTypeScore
1Qwen3.7 MaxAlibabaClosedqwen3-7-max
2DeepSeek V4 Pro (Max)DeepSeekOpendeepseek-v4-pro-max
4DeepSeek V4 Pro (High)DeepSeek · Open weight94.0%
5Qwen3.7 PlusAlibaba · Closed92.9%
6Kimi K2.6Moonshot AI · Open weight92.7%
7GLM-5.2Z.AI · Open weight92.5%
8DeepSeek V4 Flash (High)DeepSeek · Open weight91.9%
9Qwen3.5 397BAlibaba · Open weight87.9%
10Qwen3.6 PlusAlibaba · Closed87.8%
11Kimi K2.5Moonshot AI · Open weight87.1%
12GLM-5Z.AI · Open weight86.4%
13Claude Opus 4.5Anthropic · Closed85.3%
14MAI-Thinking-1Microsoft · Closed84.9%
15Qwen3.6-27BAlibaba · Open weight84.3%
16Qwen3.6-35B-A3BAlibaba · Open weight83.6%
17GLM-5.1Z.AI · Open weight82.6%
18ZAYA1-8BZyphra · Open weight71.6%
19DeepSeek V4 FlashDeepSeek · Open weight40.8%
20DeepSeek V4 ProDeepSeek · Open weight31.7%
21MiniCPM5-1BOpenBMB · Open weight25.8%