PRBench Finance: leaderboard
Professional Reasoning Bench Finance evaluates frontier LLMs on complex financial reasoning tasks including analysis, modeling, and decision-making.
Metric: Score. Source: scale.com. Status: years away from saturation. 32 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Muse Spark 1.1 | 55.01 |
| 2 | Claude Fable 5 | 53.86 |
| 3 | Claude Opus 4.6 (Non-reasoning) | 53.28 |
| 4 | Muse Spark | 52.44 |
| 5 | GPT-5 | 51.32 |
| 6 | GPT-5 Pro | 51.06 |
| 7 | Claude Fable 5.1 (xHigh) | 50.8 |
| 8 | GPT-5.6 Sol (Max) | 50.45 |
| 9 | O3 Pro | 49.08 |
| 10 | GPT-5.1 (Thinking) | 48.01 |
| 11 | O3 | 47.69 |
| 12 | Kimi K2.5 | 46.51 |
| 13 | GPT-5.2 Pro | 46.34 |
| 14 | Claude Opus 4.5 (20251101) (Thinking) | 46.16 |
| 15 | GPT-5.4 (High) | 45.63 |
Interactive version: theaggregate.ai/benchmark?slug=prbench-finance · How It Works · Data refreshed daily, snapshot 2026-09-05.