| # | Model | Provider | Type | Score | |
|---|---|---|---|---|---|
| 1 | MiMo-V2.6-Flash | Xiaomi | Open weights | 95.1% | |
| 2 | MiMo-V2.6-Pro | Xiaomi | Open weights | 94.0% | |
| 3 | DeepSeek V4.1 Flash | DeepSeek | Open weights | 88.1% | |
| 4 | Fugu Cyber | Sakana AI · Closed | 86.9% | ||
| 6 | Gemini 3.8 Flash Cyber | Google · Closed | 86.2% | ||
| 7 | Step 5 Preview | StepFun · Closed | 84.7% | ||
| 8 | GPT-5.6 Sol | OpenAI · Closed | 84.5% | ||
| 9 | GLM-5.3 | Z.AI · Open weight | 84.5% | ||
| 10 | Claude Mythos 5 | Anthropic · Closed | 83.8% | ||
| 11 | DeepSeek V4 Pro 0813 | DeepSeek · Closed | 83.3% | ||
| 12 | Gemini 3.5 Flash Cyber | Google · Closed | 83.2% | ||
| 13 | Claude Mythos Preview | Anthropic · Closed | 83.1% | ||
| 14 | GPT-5.6 Terra | OpenAI · Closed | 81.8% | ||
| 15 | GPT-5.5 | OpenAI · Closed | 81.8% | ||
| 16 | GPT-5.4 | OpenAI · Closed | 79.0% | ||
| 17 | Hy4 preview | Tencent · Open weight | 78.4% | ||
| 18 | GPT-5.6 Luna | OpenAI · Closed | 77.9% | ||
| 19 | DeepSeek V4 Flash 0731 | DeepSeek · Closed | 76.7% | ||
| 20 | Claude Opus 4.7 (Adaptive) | Anthropic · Closed | 73.1% | ||
| 21 | GLM-5.1 | Z.AI · Open weight | 68.7% | ||
| 22 | Claude Opus 4.6 | Anthropic · Closed | 66.6% | ||
| 23 | Claude Sonnet 4.6 | Anthropic · Closed | 65.2% | ||
| 24 | Muse Spark 1.1 | Meta · Closed | 59.0% | ||
| 25 | Claude Opus 4.5 | Anthropic · Closed | 50.6% | ||
| 26 | Muse Spark | Meta · Closed | 43.5% | ||
| 27 | GLM-5 | Z.AI · Open weight | 43.2% |