ForecastBench Baseline - Scratchpad - Market: leaderboard
ForecastBench Baseline leaderboard, where the ForecastBench team runs models out of the box, scratchpad prompt (step-by-step reasoning before the probability). Market component: questions from prediction markets and forecasting platforms; difficulty-adjusted Brier Index on 0-100 (higher is better; 50 equals always forecasting 0.5). Superforecaster and public medians, naive baselines and LLM crowds are excluded.
Metric: Market Brier Index (difficulty-adjusted, higher is better). Source: www.forecastbench.org. Saturation forecast: Around 2031. 36 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | O3 (2025-04-16) | 62.6 | #125 |
| 2 | Claude Opus 4.1 (20250805) | 61.2 | #139 |
| 3 | Claude 3.7 Sonnet (20250219) | 61 | #209 |
| 4 | O4 Mini (2025-04-16) | 60.7 | #183 |
| 5 | Claude 3.5 Sonnet (20241022) | 60.6 | #302 |
| 6 | GPT-4.1 | 60.2 | #258 |
| 7 | Claude Opus 4 (20250514) | 59.3 | #164 |
| 8 | GPT-4.5 (Preview) | 58.9 | #260 |
| 9 | DeepSeek R1 | 58.8 | #262 |
| 10 | Gemini 1.5 Pro | 58.6 | #454 |
| 11 | Claude Sonnet 4 (20250514) | 58.3 | #223 |
| 12 | Claude 3.5 Sonnet (20240620) | 58 | #308 |
| 13 | Gemini 2.5 Flash (Preview 04-17) | 57.9 | #246 |
| 14 | QwQ 32B-Preview | 57.9 | #671 |
| 15 | GPT-4o (2024-05-13) | 57.5 | #337 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=forecastbench-baseline-scratchpad-market · How It Works · Data refreshed daily, snapshot 2026-11-05.