Legal Benchmarks - Data Extraction - Form: leaderboard
Legal Benchmarks Data Extraction: Form is the share of applicable binary presentation checks (output usable as delivered) passed in both attempts, judged by GLM 5.3 Flash with a separate checklist that has no lawyer-labelled accuracy measurement. It does not affect the site's ranking; %, higher is better.
Metric: Form (%, binary form checks passed in both attempts). Source: www.legalbenchmarks.ai. Saturation forecast: Around April 2027. 26 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-6 Astra | 83.5 | #6 |
| 2 | GPT-6.1 Sol | 83 | #7 |
| 3 | Claude Opus 5.5 | 82.2 | #3 |
| 4 | GPT-5.6 Terra | 80.9 | #40 |
| 5 | GPT-5.6 Luna | 79.6 | #79 |
| 6 | GPT-5.6 Sol | 78.9 | #22 |
| 7 | Claude Sonnet 5.5 | 77.6 | #14 |
| 8 | Claude Haiku 5.5 | 77.4 | #43 |
| 9 | GPT-6.1 Terra | 75.4 | #15 |
| 10 | Kimi K3 | 75 | #35 |
| 11 | MiniMax-M3 | 72.8 | #151 |
| 12 | Claude Sonnet 5 | 70.6 | #65 |
| 13 | Claude Fable 5.1 | 70.6 | #10 |
| 14 | Grok 4.6 | 69.3 | #29 |
| 15 | DeepSeek V4 Pro | 67.4 | #112 |
Interactive version: theaggregate.ai/benchmark?slug=legal-benchmarks-data-extraction-form · How It Works · Data refreshed daily, snapshot 2026-11-05.