| # | Model | Provider | Type | Score | |
|---|---|---|---|---|---|
| 1 | Claude Opus 4.6 | Anthropic | Closed | opus46 | |
| 2 | Step 3.7 Flash | StepFun | Open | step_37_flash | |
| 3 | Claude Sonnet 4.6 | Anthropic | Closed | sonnet46 | |
| 4 | MiMo-V2.5-Pro | Xiaomi · Closed | 63.8% | ||
| 5 | Muse Spark | Meta · Closed | 63.8% | ||
| 6 | Kimi K2.6 | Moonshot AI · Open weight | 62.3% | ||
| 7 | MiMo-V2.5 | Xiaomi · Closed | 62.3% | ||
| 8 | GLM-5.1 | Z.AI · Open weight | 62.3% | ||
| 9 | GPT-5.4 | OpenAI · Closed | 60.3% | ||
| 10 | DeepSeek V4 Pro | DeepSeek · Open weight | 59.8% | ||
| 11 | SenseNova 6.7 Flash-Lite | SenseTime | 58.8% | ||
| 12 | Qwen3.6 Plus | Alibaba · Closed | 58.8% | ||
| 13 | Gemini 3.1 Pro | Google · Closed | 57.8% | ||
| 14 | DeepSeek V4 Flash | DeepSeek · Open weight | 57.8% | ||
| 15 | MiMo-V2-Pro | Xiaomi · Closed | 57.8% | ||
| 16 | Qwen3.5 397B | Alibaba · Open weight | 56.8% | ||
| 17 | GLM-5-Turbo | Z.AI · Closed | 55.8% | ||
| 18 | GLM-5V-Turbo | Z.AI · Closed | 53.8% | ||
| 19 | Kimi K2.5 | Moonshot AI · Open weight | 52.3% | ||
| 20 | Agnes-2.0-flash | SapiensAI | 51.8% | ||
| 21 | Gemini 3 Flash | Google · Closed | 49.2% | ||
| 22 | Mach-Mind-4-pro | Li | 49.2% | ||
| 23 | MiniMax M2.7 | MiniMax · Open weight | 48.7% | ||
| 24 | MiMo-V2-Omni | Xiaomi · Closed | 45.2% | ||
| 25 | DeepSeek V3.2 | DeepSeek · Open weight | 40.2% | ||
| 26 | Nemotron 3 Super 100B | NVIDIA · Open weight | 5.5% |