OlympiadBench: leaderboard
8,476 Olympiad-level math and physics problems from international competitions, requiring advanced multi-step reasoning.
Metric: Avg Accuracy (%). Source: github.com. Saturation forecast: Around November 2027. 7 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o | 25.89 | #343 |
| 2 | Claude 3 Opus | 7.65 | #571 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=olympiadbench · How It Works · Data refreshed daily, snapshot 2026-10-20.