DLawBench - Elicitation: leaderboard
Metric: Elicitation score (0-100), a rate scaled by 100: mean of fact coverage (annotated case facts the lawyer collects) and inquiry (expert-specified follow-up questions asked), averaged over the two jurisdictions, on 461 real Chinese-law and U.S.-law cases replayed as multi-turn consultations with a Claude Sonnet 4.6 client simulator in four narrative styles; judged by a GPT-5.1, Claude Opus 4.6 and Gemini 3.1 Pro panel (median; same-vendor judges recuse), empty memos count as zero; provider decoding defaults; higher is better. Source: arxiv.org. Saturation forecast: Around October 2027. 26 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.5 | 70.7 |
| 2 | GPT-5.4 | 68 |
| 3 | GPT-5.2 | 67.5 |
| 4 | Grok 4.1 Fast | 59.6 |
| 5 | Seed 2.0 Pro | 57 |
| 6 | Claude Opus 4.6 | 56.6 |
| 7 | Gemini 3.1 Pro (Preview) | 56.5 |
| 8 | GLM-5.1 | 56.3 |
| 9 | Qwen 3.6 Max Preview | 55.5 |
| 10 | Kimi K2.6 | 55.3 |
| 11 | DeepSeek V4 Pro | 54.5 |
| 12 | Claude Opus 4.7 | 54.3 |
| 13 | Kimi K2.5 | 53.5 |
| 14 | GLM-5 | 52.1 |
| 15 | Qwen 3.6 Plus | 51.5 |
Interactive version: theaggregate.ai/benchmark?slug=dlawbench-elicitation · How It Works · Data refreshed daily, snapshot 2026-09-29.