OTIS Mock AIME 2024-25: leaderboard
Forty-five problems from three student-written mock AIME papers in the OTIS program, each answered with one integer from 0 to 999. Epoch AI prompts the model to work step by step, replaces the three picture problems with TikZ or Asymptote code, and runs each problem eight times. An extractor model pulls the final answer and it is matched exactly against the key; higher is better.
Metric: Accuracy (%). Source: epoch.ai. Saturation forecast: Estimated already saturated. 304 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.5 (xHigh) | 100 | #31 (GPT-5.5) |
| 2 | GPT-5.6 Sol (Max) | 100 | #19 (GPT-5.6 Sol) |
| 3 | GPT-6 Astra (Max) | 100 | #6 (GPT-6 Astra) |
| 4 | GPT-6.1 Sol (Max) | 100 | #7 (GPT-6.1 Sol) |
| 5 | GPT-6 Sol (Max) | 100 | #14 (GPT-6 Sol) |
| 6 | Claude Fable 5 (High) | 100 | #16 (Claude Fable 5) |
| 7 | Claude Fable 5.1 (Max) | 100 | #9 (Claude Fable 5.1) |
| 8 | Gemini 4 Argon (High) | 100 | #8 (Gemini 4 Argon) |
| 9 | Claude Opus 5.5 (Max) | 100 | #3 (Claude Opus 5.5) |
| 10 | Claude Sonnet 5.5 (Max) | 100 | #11 (Claude Sonnet 5.5) |
| 11 | GPT-5.5 Pro (xHigh) | 100 | #17 (GPT-5.5 Pro) |
| 12 | GPT-5.6 Terra (Max) | 99.72 | #41 (GPT-5.6 Terra) |
| 13 | Claude Fable 5 (Max) | 99.72 | #16 (Claude Fable 5) |
| 14 | Muse Wonder (Max) | 99.44 | #15 (Muse Wonder) |
| 15 | Qwen 3.8 Max (xHigh) | 99.44 | #35 (Qwen 3.8 Max) |
No result here: #10 Muse Wonder Full.
Interactive version: theaggregate.ai/benchmark?slug=otis-mock-aime-2024-25 · How It Works · Data refreshed daily, snapshot 2026-10-20.