context.vn
Menu

cyber Gym

26 models evaluated

#ModelProviderTypeScore
1MiMo-V2.6-FlashXiaomiOpen weights95.1%
2MiMo-V2.6-ProXiaomiOpen weights94.0%
3DeepSeek V4.1 FlashDeepSeekOpen weights88.1%
4Fugu CyberSakana AI · Closed86.9%
6Gemini 3.8 Flash CyberGoogle · Closed86.2%
7Step 5 PreviewStepFun · Closed84.7%
8GPT-5.6 SolOpenAI · Closed84.5%
9GLM-5.3Z.AI · Open weight84.5%
10Claude Mythos 5Anthropic · Closed83.8%
11DeepSeek V4 Pro 0813DeepSeek · Closed83.3%
12Gemini 3.5 Flash CyberGoogle · Closed83.2%
13Claude Mythos PreviewAnthropic · Closed83.1%
14GPT-5.6 TerraOpenAI · Closed81.8%
15GPT-5.5OpenAI · Closed81.8%
16GPT-5.4OpenAI · Closed79.0%
17Hy4 previewTencent · Open weight78.4%
18GPT-5.6 LunaOpenAI · Closed77.9%
19DeepSeek V4 Flash 0731DeepSeek · Closed76.7%
20Claude Opus 4.7 (Adaptive)Anthropic · Closed73.1%
21GLM-5.1Z.AI · Open weight68.7%
22Claude Opus 4.6Anthropic · Closed66.6%
23Claude Sonnet 4.6Anthropic · Closed65.2%
24Muse Spark 1.1Meta · Closed59.0%
25Claude Opus 4.5Anthropic · Closed50.6%
26Muse SparkMeta · Closed43.5%
27GLM-5Z.AI · Open weight43.2%