Benchmarks / Reasoning & Math
OTIS Mock AIME
Mock AIME 2024–2025 problems evaluated by Epoch AI.
maintained by Epoch AI · 154 observations on record
Versions
1 version
| Version | Tasks |
|---|---|
| OTIS Mock AIME 2024–2025 current otis-mock-aime | — |
Current leaderboard
OTIS Mock AIME 2024–2025 · Accuracy · top 25 of 154 systems
| # | System | Accuracy % |
|---|---|---|
| 1 | GPT 5.6 Solmax effort OpenAI | 100.0% |
| 2 | GPT 5.5 Pro Pre Releasexhigh effort OpenAI | 100.0% |
| 3 | GPT 5.5 Pre Releasexhigh effort OpenAI | 100.0% |
| 4 | Claude Fable 5max effort Anthropic | 99.7% |
| 5 | GPT 5.6 Terramax effort OpenAI | 99.7% |
| 6 | Claude Opus 4.8max effort Anthropic | 98.3% |
| 7 | GPT 5.6 Lunamax effort OpenAI | 98.3% |
| 8 | Claude Opus 4.7xhigh effort Anthropic | 97.8% |
| 9 | Grok 4.5high effort xAI | 97.8% |
| 10 | Deepseek V4 Promax effort DeepSeek | 96.7% |
| 11 | Kimi K2 7 Code Moonshot | 96.4% |
| 12 | GPT 5.2 (2025-12-11)high effort OpenAI | 96.1% |
| 13 | Kimi K2 6 Moonshot AI | 96.1% |
| 14 | GPT 5.2 (2025-12-11)xhigh effort OpenAI | 96.1% |
| 15 | Gemini 3.1 Pro Preview Google DeepMind | 95.6% |
| 16 | Gemini 3.5 Flashhigh effort Google DeepMind | 95.6% |
| 17 | GPT 5.4 (2026-03-05)xhigh effort OpenAI | 95.3% |
| 18 | Qwen3 7max effort Alibaba | 95.0% |
| 19 | Claude Sonnet 5xhigh effort Anthropic | 94.7% |
| 20 | Claude Opus 4.664K budget Anthropic | 94.4% |
| 21 | GPT 5.2 (2025-12-11)medium effort OpenAI | 93.9% |
| 22 | Grok 4.3high effort xAI | 93.3% |
| 23 | Claude Opus 4.632K budget Anthropic | 93.1% |
| 24 | Gemini 3 Flash Preview Google DeepMind | 92.8% |
| 25 | Grok 4.20 0309 Reasoning xAI | 92.2% |
One row per evaluated system — reasoning-effort variants rank separately.