SciPredict
Forecasting · 2026-01-15
Forecast accuracy on real scientific experiment outcomes in physics, biology, and chemistry.
Top models (higher is better)
| Model | Score |
|---|---|
| Gemini 3 Pro Preview | 25.3 |
| Opus 4.5 | 23.1 |
| Sonnet 4.5 | 22.6 |
| Opus 4.1 | 22.2 |
| Gemini 3 Flash Preview | 22.2 |
| GPT-5.2 | 20.6 |
| o3-mini | 19.8 |
| DeepSeek-V3 | 19.2 |
| Llama 3.3 70B Instruct | 18.2 |
| o3 | 17.9 |
| Gemini 2.5 Pro | 17.0 |
| Qwen3 32B | 17.0 |
| Qwen3-235B-A22B | 16.6 |
| O4 Mini | 16.2 |
| Llama 3.1 8B Instruct | 14.7 |