Models Tested
12
Best Score
96.7%
Median Score
85%
Scoring: accuracy
Introduced: 2024-02
Maintainer: Mathematical Association of America
Leaderboard (12 models)
| # | Model | Developer | Score |
|---|---|---|---|
| 🥇 | o3 | OpenAI | 96.7% |
| 🥈 | o4-mini | OpenAI | 93.4% |
| 🥉 | Gemini 2.5 Pro | Google DeepMind | 92% |
| 4 | o3-mini | OpenAI | 87.3% |
| 5 | Grok-3 | xAI | 86.7% |
| 6 | Claude Opus 4 | Anthropic | 86.7% |
| 7 | Claude Sonnet 4 | Anthropic | 83.3% |
| 8 | Claude 3.7 Sonnet | Anthropic | 80% |
| 9 | DeepSeek R1 | DeepSeek | 79.8% |
| 10 | o1 | OpenAI | 74.4% |
| 11 | o1-preview | OpenAI | 44.6% |
| 12 | DeepSeek V3 | DeepSeek | 39.2% |