FrontierScience Olympiad
Science · 2025-12-16
FrontierScience Olympiad is the Olympiad-style track of the FrontierScience scientific reasoning benchmark. Scores are reported as accuracy-style percentage points, with higher scores indicating stronger performance.
Top models (higher is better)
| Model | Score |
|---|---|
| Claude Mythos Preview | 83.0 |
| Opus 4.8 | 79.5 |
| GPT-5.5 | 79.0 |
| GLM-5.2 | 74.5 |
| GPT-5.4 Mini | 74.0 |
| Opus 4.6 | 72.0 |