SEAL - Professional Reasoning Benchmark - Legal: leaderboard
500 expert-curated legal questions spanning 114 countries and 47 US jurisdictions, measuring legal accuracy, reasoning transparency, ethical risk disclosure, and handling of ambiguous scenarios.
Metric: Score. Source: scale.com. Saturation forecast: Around 2031. 42 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Muse Spark 1.3 | 61.56 | #24 |
| 2 | Muse Wonder | 58.4 | #15 |
| 3 | Muse Spark 1.1 | 57.05 | #52 |
| 4 | Claude Fable 5 | 52.56 | #16 |
| 5 | Muse Spark | 52.29 | #64 |
| 6 | Claude Opus 4.6 (Non-reasoning) | 52.27 | #63 (Claude Opus 4.6) |
| 7 | Claude Fable 5.1 | 51.6 | #9 |
| 8 | GPT-5.6 Sol (Max) | 50.5 | #19 (GPT-5.6 Sol) |
| 9 | GPT-5 Pro | 49.89 | #81 |
| 10 | O3 Pro | 49.67 | #116 |
| 11 | Gemini 4 Argon | 49.41 | #8 |
| 12 | GPT-5.1 (Thinking) | 49.33 | #135 (GPT-5.1) |
| 13 | GPT-5 | 48.96 | #98 |
| 14 | Gemini 3.8 Flash | 48.9 | #21 |
| 15 | O3 | 48.57 | #124 |
No result here: #10 Muse Wonder Full.
Interactive version: theaggregate.ai/benchmark?slug=seal-professional-reasoning-benchmark-legal · How It Works · Data refreshed daily, snapshot 2026-10-20.