LegalCiteBench - Case Matching: 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. In case matching the model identifies the source case of an anonymized legal scenario; judge score from 0 to 100, higher is better.

Metric: Cat4-1 judge score (0-100) for matching an anonymized legal scenario to its source case name and citation (1,997 instances), 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; higher is better. Source: arxiv.org. Saturation forecast: Not forecast. 21 models tracked.

Top models

#ModelScoreOverall rank
1GPT-5 Mini74#192
2DeepSeek-R1-Distill-Qwen-7B55.53#1400
3DeepSeek R1 Distill Qwen 1.5B53.42#1522
4Qwen 3 14B43.26#542
5Qwen 2.5 7B43.24#898
6O4 Mini42.39#185
7DeepSeek V3.142.29#276
8Claude Sonnet 4.541.64#148
9Gemini 2.5 Flash41.12#254
10Claude Haiku 4.540.56#288
11Phi-440.32#708
12Llama 3.1 8B40.24#1131
13Qwen 3 4B40.05#835
14GPT-4o Mini39.54#599
15Qwen 3 30B A3B39.32#499

No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #6 GPT-6.1 Sol.

Interactive version: theaggregate.ai/benchmark?slug=legalcitebench-case-matching · How It Works · Data refreshed daily, snapshot 2026-11-06.