CUSP Science — leaderboard
Forecasting benchmark for scientific progress where models answer cutoff-conditioned binary, multiple-choice, date, and free-response questions about future scientific milestones.
Metric: FRQ Pass (%). Source: seanwu25.github.io. Status: years away from saturation. 6 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.4 | 60.3 |
| 2 | DeepSeek R1 | 20 |
| 3 | Claude Sonnet 4.5 | 14 |
| 4 | GPT-OSS-20B | 11.9 |
| 5 | GPT-4o | 3.9 |
| 6 | Llama 3.3 70B | 2.5 |
Interactive version: theaggregate.ai/benchmark?slug=cusp-science · How the rankings work · Data refreshed daily, snapshot 2026-07-22.