LexGenius
Models answer Chinese legal questions drawn from cases and exams. The authors compare results across areas including legal reasoning, ethics, and judicial practice.
- Made by
- Wenjin Liu and colleagues
- Released
- November 2025
- Kind
- Law
- Level
- University
- Scored
- Checked automatically
- Answers
- Multiple choice
- Tasks
- All public
- Measures
- Capability
- Subject
- Law
- Language
- Chinese
- Interaction
- Single turn
- Items from
- Real work and use, Exams and contests
- Refresh
- One fixed set
- Run by
- Anyone
- In views
- Law and finance
- Strongest model tested
-
Qwen3-8B #154 on the capability index
- Last model added
- April 2025
- Scores refreshed
- November 20th, 2026
DeepSeek-R1 (snapshot unspecified) tops the leaderboard with 64.6% Qwen2.5-7B leads with 59.7% among models released by April 2025
Best score per entry from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. The best model we track is Qwen2.5-7B, at 59.7%. This view shows only models we track, each at its best.
The newest model with a score here came out in April 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
- 1 DeepSeek-R1 (snapshot unspecified) Chain of thought 64.6%
- 2 DeepSeek-V3 (snapshot unspecified) Chain of thought 64.4%
- 3
Qwen2.5-7B Naive prompt 59.7%
- 4
Qwen3-8B Naive prompt 55.9%
- 5
Qwen3-4B Naive prompt 53.6%
- 6 GLM-4-9B Chain of thought 53.5%
- 7 GPT-4.1 nano Chain of thought 53.4%
- 8 GPT-4o mini Chain of thought 52.6%
- 9
Qwen2.5-1.5B Chain of thought 48.2%
- 10 LLaMA-3.2-8B Chain of thought 43.1%
- 11 DeepSeek-7B Naive prompt 38.6%
- 12
Llama 3.2 1B Naive prompt 35.3%
Source: The LexGenius paper
- 1
Qwen2.5-7B Naive prompt 59.7%
- 2
Qwen3-8B Naive prompt 55.9%
- 3
Qwen3-4B Naive prompt 53.6%
- 4 GLM-4-9B Chain of thought 53.5%
- 5 GPT-4.1 nano Chain of thought 53.4%
- 6 GPT-4o mini Chain of thought 52.6%
- 7
Qwen2.5-1.5B Chain of thought 48.2%
- 8
Llama 3.2 1B Naive prompt 35.3%
Source: The LexGenius paper
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | DeepSeek-R1 (snapshot unspecified) | 64.6% | Chain of thought | Official leaderboard | |
| 2 | DeepSeek-V3 (snapshot unspecified) | 64.4% | Chain of thought | Official leaderboard | |
| 3 | | 59.7% | Naive prompt | Sep 2024 | Official leaderboard |
| 4 | | 55.9% | Naive prompt | Apr 2025 | Official leaderboard |
| 5 | | 53.6% | Naive prompt | Apr 2025 | Official leaderboard |
| 6 | GLM-4-9B | 53.5% | Chain of thought | Jun 2024 | Official leaderboard |
| 7 | GPT-4.1 nano | 53.4% | Chain of thought | Apr 2025 | Official leaderboard |
| 8 | GPT-4o mini | 52.6% | Chain of thought | Jul 2024 | Official leaderboard |
| 9 | | 48.2% | Chain of thought | Sep 2024 | Official leaderboard |
| 10 | LLaMA-3.2-8B | 43.1% | Chain of thought | Official leaderboard | |
| 11 | DeepSeek-7B | 38.6% | Naive prompt | Official leaderboard | |
| 12 | | 35.3% | Naive prompt | Sep 2024 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 59.7% | Naive prompt | Sep 2024 | Official leaderboard |
| 2 | | 55.9% | Naive prompt | Apr 2025 | Official leaderboard |
| 3 | | 53.6% | Naive prompt | Apr 2025 | Official leaderboard |
| 4 | GLM-4-9B | 53.5% | Chain of thought | Jun 2024 | Official leaderboard |
| 5 | GPT-4.1 nano | 53.4% | Chain of thought | Apr 2025 | Official leaderboard |
| 6 | GPT-4o mini | 52.6% | Chain of thought | Jul 2024 | Official leaderboard |
| 7 | | 48.2% | Chain of thought | Sep 2024 | Official leaderboard |
| 8 | | 35.3% | Naive prompt | Sep 2024 | Official leaderboard |
Rankings by part
We read 7 parts from the benchmark’s paper, each ranking models on one part of the test, such as Legal Understanding, Legal Reasoning and Legal Application. Qwen2.5-7B leads every one of them. Arrows show how far a model moved from its overall rank among the models we track.
Legal Understanding
8 models we track have a score here.
- 1
Qwen2.5-7B 61.4%
- 2
Qwen3-8B 58.8%
- 3 GPT-4.1 nano ↑2 54.6%
- 4
Qwen3-4B ↓1 53.2%
- 5 GLM-4-9B ↓1 53%
- 6 GPT-4o mini 49.8%
- 7
Qwen2.5-1.5B 44%
- 8
Llama 3.2 1B 29.4%
Legal Reasoning
8 models we track have a score here.
- 1
Qwen2.5-7B 57.4%
- 2
Qwen3-8B 52.7%
- 3
Qwen3-4B 50.6%
- 4 GLM-4-9B 47.8%
- 5 GPT-4o mini ↑1 43.9%
- 6 GPT-4.1 nano ↓1 43.9%
- 7
Qwen2.5-1.5B 42.3%
- 8
Llama 3.2 1B 25.8%
Legal Application
8 models we track have a score here.
- 1
Qwen2.5-7B 45.3%
- 2 GPT-4.1 nano ↑3 40%
- 3
Qwen3-4B 39%
- 4
Qwen3-8B ↓2 38.7%
- 5 GPT-4o mini ↑1 37%
- 6 GLM-4-9B ↓2 36.5%
- 7
Qwen2.5-1.5B 35.8%
- 8
Llama 3.2 1B 26.5%
Legal Ethics
8 models we track have a score here.
- 1
Qwen2.5-7B 64%
- 2
Qwen3-8B 62.9%
- 3 GPT-4o mini ↑3 62.2%
- 4 GLM-4-9B 61.4%
- 5
Qwen3-4B ↓2 61%
- 6 GPT-4.1 nano ↓1 59.6%
- 7
Qwen2.5-1.5B 56%
- 8
Llama 3.2 1B 44.4%
Legal Language
8 models we track have a score here.
- 1
Qwen2.5-7B 66.1%
- 2 GLM-4-9B ↑2 63.2%
- 3
Qwen3-8B ↓1 63.1%
- 4 GPT-4.1 nano ↑1 62.7%
- 5
Qwen3-4B ↓2 57.6%
- 6 GPT-4o mini 57.5%
- 7
Qwen2.5-1.5B 53.3%
- 8
Llama 3.2 1B 39.9%
Law and Society
8 models we track have a score here.
- 1
Qwen2.5-7B 66.3%
- 2 GPT-4o mini ↑4 65.8%
- 3
Qwen3-4B 62.2%
- 4 GLM-4-9B 61.9%
- 5
Qwen3-8B ↓3 61.2%
- 6 GPT-4.1 nano ↓1 60.5%
- 7
Qwen2.5-1.5B 59.8%
- 8
Llama 3.2 1B 50.3%
Judicial Practice
8 models we track have a score here.
- 1
Qwen2.5-7B 57.6%
- 2
Qwen3-8B 55.4%
- 3 GPT-4o mini ↑3 52.6%
- 4 GPT-4.1 nano ↑1 52.6%
- 5
Qwen3-4B ↓2 51.8%
- 6 GLM-4-9B ↓2 51%
- 7
Qwen2.5-1.5B 46.4%
- 8
Llama 3.2 1B 33%
Scores at each effort level
Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.
Source: The LexGenius paper
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| Naive prompt | 59.7% | ||||
| Chain of thought | 59.4% | ||||
| Naive prompt | 55.9% | ||||
| Chain of thought | 55.3% | ||||
| Naive prompt | 53.6% | ||||
| GLM-4-9B | Chain of thought | 53.5% | |||
| GPT-4.1 nano | Chain of thought | 53.4% | |||
| GLM-4-9B | Naive prompt | 53.2% | |||
| Chain of thought | 53.2% | ||||
| GPT-4o mini | Chain of thought | 52.6% | |||
| GPT-4o mini | Naive prompt | 52.5% | |||
| Chain of thought | 48.2% | ||||
| GPT-4.1 nano | Naive prompt | 48.2% | |||
| Naive prompt | 47.6% | ||||
| Naive prompt | 35.3% | ||||
| Chain of thought | 34.9% |
Scores at each effort level
Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.
Source: The LexGenius paper
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| Naive prompt | 59.7% | ||||
| Chain of thought | 59.4% | ||||
| Naive prompt | 55.9% | ||||
| Chain of thought | 55.3% | ||||
| Naive prompt | 53.6% | ||||
| GLM-4-9B | Chain of thought | 53.5% | |||
| GPT-4.1 nano | Chain of thought | 53.4% | |||
| GLM-4-9B | Naive prompt | 53.2% | |||
| Chain of thought | 53.2% | ||||
| GPT-4o mini | Chain of thought | 52.6% | |||
| GPT-4o mini | Naive prompt | 52.5% | |||
| Chain of thought | 48.2% | ||||
| GPT-4.1 nano | Naive prompt | 48.2% | |||
| Naive prompt | 47.6% | ||||
| Naive prompt | 35.3% | ||||
| Chain of thought | 34.9% |
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: The LexGenius paper
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official LexGenius leaderboard
No lab model card reports it yet
We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.
More in Law, finance and medicine
- MedXpertQA-MM In use Models answer expert-level medical exam questions that combine images with clinical records and test results. Scoring is accuracy.
Qwen 3.8 Max 80.4%
- HealthBench Professional In use 525 physician-authored tasks from real clinician chats (care consult, documentation, research) with physician-adjudicated rubrics. Claude Sonnet 5.5 69.2%
- HealthBench In use 5,000 realistic health conversations graded against physician-written rubrics. A length-adjusted variant penalizes verbosity. GPT-5 67.2%
- FinSearchComp In use Agents search the web to fetch live financial data, look up history, and run multi-step investigations, across 635 expert-written questions. Scored by answer accuracy.
Ling-3.1-flash 81.7%
- KorMedMCQA In use Models answer 7,469 multiple-choice questions from Korean licensing exams for doctors, nurses, pharmacists and dentists, held from 2012 to 2024. o1-preview 92.7%
- Harvey LAB-AA v1.1 In use Artificial Analysis's run of Harvey's Legal Agent Benchmark on Harvey's newer private set of 120 tasks across 24 practice areas. A deliverable with a material hallucination scores zero, and the headline is the share of judges who find every rubric criterion passed. GPT-6.2 Astra 15%
Data audit
9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 19 months ago, in April 2025, for Qwen3-4B. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 1 warning 3 pass 1 note
- Pass Independent runs Scores from the benchmark’s paper (12 entries).
- Pass Models scored 8 models we track, plus 4 other entries on its leaderboard.
- Warning Newest score The newest came 19 months ago, in April 2025, for Qwen3-4B.
- Note Frontier models None of the top 10 has a score. The strongest tested is Qwen3-8B (#154).
- Pass Headroom The best is 59.7%, under 85%, so it still separates models.
Leaderboard import 3 pass
- Pass Importer We read the board on arxiv.org: 16 rows for models we track and 4 other entries at the last read.
- Pass Last read November 20th, 2026.
- Pass Read worked The last read worked.
Consistency 1 pass 1 note
- Note Lab claims No model has both a lab number and an independent run.
- Pass Known issues No published audit found problems.