IMO-AnswerBench: leaderboard
International Mathematical Olympiad answer benchmark evaluating final-answer correctness on high-difficulty olympiad-style mathematical problems.
Metric: Score (%). Source: llm-stats.com. Status: years away from saturation. 20 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Nemotron 3 Ultra | 92.3 |
| 2 | GLM-5.2 | 91 |
| 3 | Qwen 3.7 Max | 90 |
| 4 | Hy3 | 90 |
| 5 | DeepSeek V4 Pro (Max) | 89.8 |
| 6 | DeepSeek V4 Flash (Max) | 88.4 |
| 7 | Kimi K2.6 | 86 |
| 8 | Qwen 3.7 Plus | 86 |
| 9 | Step 3.5 Flash | 85.4 |
| 10 | DeepSeek V4 Flash (0423) | 85.1 |
| 11 | Qwen 3.6 Plus | 83.8 |
| 12 | GLM-5.1 | 83.8 |
| 13 | GLM-4.7 | 82 |
| 14 | Kimi K2.5 | 81.8 |
| 15 | Qwen 3.5 397B A17B | 80.9 |
Interactive version: theaggregate.ai/benchmark?slug=imo-answerbench · How It Works · Data refreshed daily, snapshot 2026-09-05.