Vendors
Model creators/vendors and how their models compare across the benchmark.
| Vendor | Models | Avg Score | Best Score ▼ | Best Model |
|---|---|---|---|---|
| Anthropic | 21 | 90.66% | 96.43% | Claude Opus 5.5 (Reasoning) |
| Qwen | 20 | 88.01% | 96.21% | Qwen 3.8 Max (Reasoning, XHigh) |
| 26 | 85.74% | 95.72% | Gemini 3.8 Flash (Reasoning, Medium) | |
| Z.AI | 12 | 89.29% | 95.20% | Z.AI GLM 5.3 (Reasoning, Max) |
| OpenAI | 38 | 86.39% | 95.15% | GPT-5.6 Sol (Reasoning, Medium) |
| tencent | 1 | 94.80% | 94.80% | Hy4 Preview (Reasoning, High) |
| meta | 5 | 92.57% | 94.71% | Muse Spark 1.2 (Reasoning, Medium) |
| DeepSeek | 12 | 85.43% | 94.30% | DeepSeek V4.1 Flash (Reasoning, High) |
| xAI | 8 | 88.94% | 94.12% | Grok 4.5 (Reasoning, High) |
| MoonshotAI | 4 | 92.12% | 92.99% | Kimi K3 (Reasoning, High) |
| aion-labs | 5 | 88.68% | 92.36% | Aion 3.5 Mini (Reasoning, High) |
| xiaomi | 4 | 87.20% | 91.50% | Xiaomi MIMO v2.6 Pro |
| thinkingmachines | 3 | 84.04% | 90.78% | Thinking Machines Inkling (Reasoning) |
| minimax | 3 | 87.80% | 90.45% | MiniMax M3 |
| bytedance-seed | 4 | 82.62% | 89.59% | ByteDance Seed 1.6 |
| Mistral AI | 13 | 73.20% | 85.32% | Mistral Medium 3.5 (Reasoning) |
| inception | 1 | 81.99% | 81.99% | Inception Mercury 2 |
| NVIDIA | 2 | 78.10% | 81.69% | Nemotron 3 Super |
| Nous Research | 2 | 75.27% | 80.80% | Hermes 3 405B |
| poolside | 2 | 77.70% | 79.99% | Laguna S 2.1 |
| Writer | 1 | 78.11% | 78.11% | Writer: Palmyra X5 |
| Meta | 1 | 77.41% | 77.41% | Llama 3.1 70B |
| TheDrummer | 1 | 72.68% | 72.68% | Cydonia 24B V4.1 |
| Microsoft | 1 | 71.45% | 71.45% | WizardLM 2 8x22b |
| arcee-ai | 1 | 67.68% | 67.68% | Arcee AI: Trinity Mini |
| Cohere | 1 | 67.04% | 67.04% | Cohere Command R+ (Aug. 2024) |