Leaderboard bars
20 ranked models · higher is better
View accessible chart data
| Model | Rank | Provider | Score |
|---|---|---|---|
| GPT-5.2 | #1 | OpenAI | 99% |
| GPT-5 Codex | #2 | OpenAI | 98.7% |
| MiMo-V2-Flash | #3 | Xiaomi | 96.3% |
| Kimi K2.5 | #4 | Moonshot AI | 96.1% |
| GLM-4.7 | #5 | Z.AI | 95% |
| Kimi K2 Thinking | #6 | Moonshot AI | 94.7% |
| GPT-5 | #7 | OpenAI | 94.3% |
| Grok 4 | #8 | xAI | 92.7% |
| DeepSeek V3.2 | #9 | DeepSeek | 92% |
| o4-mini | #10 | OpenAI | 90.7% |
| DeepSeek V3.1 Terminus | #11 | DeepSeek | 89.7% |
| Grok 4 Fast | #12 | xAI | 89.7% |
| Nova 2 Pro | #13 | Amazon | 89% |
| o3 | #14 | OpenAI | 88.3% |
| Claude Sonnet 4.5 | #15 | Anthropic | 88% |
| Gemini 2.5 Pro | #16 | 87.7% | |
| GLM-4.6 | #17 | Z.AI | 86% |
| GPT-5 mini | #18 | OpenAI | 85% |
| Grok 3 mini | #19 | xAI | 84.7% |
| MiniMax M2.1 | #20 | MiniMax | 82.7% |

