LegalCiteBench - Misleading Answer Rate: leaderboard
LegalCiteBench (2026) tests whether LLMs, without retrieval, can recover and verify the case authorities cited in 1,000 U.S. judicial opinions from the Case Law Access Project. The misleading answer rate is the share of low-scoring retrieval and matching responses that still assert a concrete citation or case rather than abstaining, in percent; lower is better.
Metric: Misleading Answer Rate (%): share of low-scoring responses (task score at most 40) on Cat1, Cat2 and Cat4-1 that still give a concrete citation or case instead of abstaining, closed-book (no retrieval) on instances derived from 1,000 U.S. judicial opinions of the Case Law Access Project, greedy decoding at temperature 0 with at most 400 generated tokens, GPT-4o-mini judge; lower is better. Source: arxiv.org. Saturation forecast: Around 2037. 21 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 Mini | 89.24 | #364 |
| 2 | DeepSeek-R1-Distill-Qwen-7B | 94.21 | #8055 |
| 3 | DeepSeek R1 Distill Qwen 1.5B | 94.62 | #9215 |
| 4 | Claude Sonnet 4.5 | 96.79 | #261 |
| 5 | O4 Mini | 96.85 | #368 |
| 6 | Qwen 3 14B | 96.97 | #1269 |
| 7 | Qwen 2.5 7B | 97.53 | #2929 |
| 8 | DeepSeek V3.1 | 97.58 | #511 |
| 9 | SaulLM-54B | 97.69 | |
| 10 | Gemini 2.5 Flash | 98.21 | #463 |
| 11 | Claude Haiku 4.5 | 98.31 | #583 |
| 12 | Qwen 3 30B A3B | 98.35 | #1156 |
| 13 | Llama 3.1 70B | 98.47 | #1499 |
| 14 | Mistral 7B | 98.69 | #7794 |
| 15 | Qwen 3 4B | 99.01 | #2348 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=legalcitebench-misleading-answer-rate · How It Works · Data refreshed daily, snapshot 2026-11-21.