LawBench
Models answer Chinese legal questions across 20 tasks, including statute recall and case prediction. Each task has its own scoring rule.
GPT-4 led the September 2023 comparison with averages of 52.35 zero-shot and 53.85 one-shot. The authors also tested GPT-3.5 Turbo, StableBeluga2, Qwen, and InternLM.
- Made by
- Zhiwei Fei and colleagues
- Released
- September 2023
- Kind
- Law
- Level
- University
- Scored
- Checked automatically
- Answers
- Short answer
- Tasks
- All public
- Measures
- Capability
- Subject
- Law
- Language
- Chinese
- Interaction
- Single turn
- Refresh
- One fixed set
- Run by
- Anyone
- In views
- Law and finance
- Strongest model tested
- GPT-3.5 Turbo Jun 2023 #209 on the capability index
- Last model added
- September 2023
- Scores refreshed
- November 20th, 2026
GPT4 (snapshot unspecified) tops the leaderboard with 53.9 GPT-3.5 Turbo leads with 44.5 among models released by September 2023
Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. 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 GPT-3.5 Turbo, at 44.5. This view shows only models we track, each at its best.
The newest model with a score here came out in September 2023, 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
- 53.9 GPT4 (snapshot unspecified)
- 44.5 GPT-3.5 Turbo
- 39.2 freewilly2_70b-hf
- 39.0 qwen-7b-chat-hf
- 37.3 internlm-chat-7b-8k-hf
- 36.1 internlm-chat-7b-hf
- 34.5 yulan-chat-2-13b-fp16-hf
- 33.0 fuzi-mingcha-7b-hf
- 32.8 chatlaw-13b-hf
- 31.4 wisdom-interrogatory-7b-hf
- 30.8 belle-2-13b-hf
- 29.7 hanfei-1.0-7b-hf
- 29.6 baichuan-13b-chat-hf
- 28.8 lexilaw-6b-hf
- 26.1 chatlaw-33b-hf
- 25.9 tigerbot-sft-7b-hf
- 25.3 lawyer-llama-13b-hf
- 24.0 baichuan-13b-base-hf
- 22.9 chatglm2-6b
- 22.5 internlm-7b
- 21.7 tigerbot-base-7b-hf
- 21.2 baichuan-7b-hf
- 20.8
Llama 2-70B
- 19.6
Qwen-7B
- 19.2 chinese-alpaca-2-7b-hf
- 19.1 vicuna-33b-hf
- 18.8 ziya-llama-13b-hf
- 18.4
Llama 2-13B
- 18.3
Llama 2-7B
- 17.7
LLaMA-65B
- 17.1 llama-2-chinese-13b-hf
- 16.6
LLaMA-13B
- 16.5 llama-30b
- 16.2 gogpt-7b
- 15.8 vicuna-13b-hf
- 15.7 moss-moon-003-base-hf
- 15.6 alpaca-7b-hf
- 15.6 moss-moon-003-sft-hf
- 14.8
LLaMA-7B
- 14.7 llama-2-70b-chat
- 14.6 llama-2-13b-chat
- 14.5 wizardlm-7b-hf
- 14.3
MPT-7B
- 13.7 xverse-13b-hf
- 13.7 vicuna-7b-hf
- 12.6 mpt-instruct-7b-hf
- 12.6 chinese-llama-2-7b-hf
- 11.6 lawgpt-7b-beta1.1-hf
- 10.7 llama-2-chinese-7b-hf
- 7.16 llama-2-7b-chat
- 2.33 lawgpt-7b-beta1.0-hf
- 1 GPT4 (snapshot unspecified) one-shot 53.9
- 2 GPT-3.5 Turbo Jun 2023, one-shot 44.5
- 3 freewilly2_70b-hf zero-shot 39.2
- 4 qwen-7b-chat-hf one-shot 39.0
- 5 internlm-chat-7b-8k-hf one-shot 37.3
- 6 internlm-chat-7b-hf one-shot 36.1
- 7 yulan-chat-2-13b-fp16-hf one-shot 34.5
- 8 fuzi-mingcha-7b-hf zero-shot 33.0
- 9 chatlaw-13b-hf zero-shot 32.8
- 10 wisdom-interrogatory-7b-hf zero-shot 31.4
- 11 belle-2-13b-hf one-shot 30.8
- 12 hanfei-1.0-7b-hf zero-shot 29.7
- 13 baichuan-13b-chat-hf one-shot 29.6
- 14 lexilaw-6b-hf zero-shot 28.8
- 15 chatlaw-33b-hf zero-shot 26.1
- 16 tigerbot-sft-7b-hf zero-shot 25.9
- 17 lawyer-llama-13b-hf zero-shot 25.3
- 18 baichuan-13b-base-hf one-shot 24.0
- 19 chatglm2-6b one-shot 22.9
- 20 internlm-7b one-shot 22.5
- 21 tigerbot-base-7b-hf zero-shot 21.7
- 22 baichuan-7b-hf one-shot 21.2
- 23
Llama 2-70B one-shot 20.8
- 24
Qwen-7B zero-shot 19.6
- 25 chinese-alpaca-2-7b-hf one-shot 19.2
- 26 vicuna-33b-hf one-shot 19.1
- 27 ziya-llama-13b-hf one-shot 18.8
- 28
Llama 2-13B one-shot 18.4
- 29
Llama 2-7B one-shot 18.3
- 30
LLaMA-65B one-shot 17.7
- 31 llama-2-chinese-13b-hf one-shot 17.1
- 32
LLaMA-13B one-shot 16.6
- 33 llama-30b one-shot 16.5
- 34 gogpt-7b one-shot 16.2
- 35 vicuna-13b-hf one-shot 15.8
- 36 moss-moon-003-base-hf one-shot 15.7
- 37 alpaca-7b-hf one-shot 15.6
- 38 moss-moon-003-sft-hf one-shot 15.6
- 39
LLaMA-7B one-shot 14.8
- 40 llama-2-70b-chat zero-shot 14.7
- 41 llama-2-13b-chat zero-shot 14.6
- 42 wizardlm-7b-hf one-shot 14.5
- 43
MPT-7B one-shot 14.3
- 44 xverse-13b-hf one-shot 13.7
- 45 vicuna-7b-hf one-shot 13.7
- 46 mpt-instruct-7b-hf one-shot 12.6
- 47 chinese-llama-2-7b-hf one-shot 12.6
- 48 lawgpt-7b-beta1.1-hf one-shot 11.6
- 49 llama-2-chinese-7b-hf one-shot 10.7
- 50 llama-2-7b-chat zero-shot 7.16
- 51 lawgpt-7b-beta1.0-hf one-shot 2.33
Source: Official LawBench leaderboard
- 1 GPT-3.5 Turbo Jun 2023, one-shot 44.5
- 2 chatglm2-6b one-shot 22.9
- 3 internlm-7b one-shot 22.5
- 4
Llama 2-70B one-shot 20.8
- 5
Qwen-7B zero-shot 19.6
- 6
Llama 2-13B one-shot 18.4
- 7
Llama 2-7B one-shot 18.3
- 8
LLaMA-65B one-shot 17.7
- 9
LLaMA-13B one-shot 16.6
- 10
LLaMA-7B one-shot 14.8
- 11
MPT-7B one-shot 14.3
Source: Official LawBench leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT4 (snapshot unspecified) | 53.9 | one-shot | Official leaderboard | |
| 2 | GPT-3.5 Turbo Jun 2023 | 44.5 | one-shot | Jun 2023 | Official leaderboard |
| 3 | freewilly2_70b-hf | 39.2 | zero-shot | Official leaderboard | |
| 4 | qwen-7b-chat-hf | 39.0 | one-shot | Official leaderboard | |
| 5 | internlm-chat-7b-8k-hf | 37.3 | one-shot | Official leaderboard | |
| 6 | internlm-chat-7b-hf | 36.1 | one-shot | Official leaderboard | |
| 7 | yulan-chat-2-13b-fp16-hf | 34.5 | one-shot | Official leaderboard | |
| 8 | fuzi-mingcha-7b-hf | 33.0 | zero-shot | Official leaderboard | |
| 9 | chatlaw-13b-hf | 32.8 | zero-shot | Official leaderboard | |
| 10 | wisdom-interrogatory-7b-hf | 31.4 | zero-shot | Official leaderboard | |
| 11 | belle-2-13b-hf | 30.8 | one-shot | Official leaderboard | |
| 12 | hanfei-1.0-7b-hf | 29.7 | zero-shot | Official leaderboard | |
| 13 | baichuan-13b-chat-hf | 29.6 | one-shot | Official leaderboard | |
| 14 | lexilaw-6b-hf | 28.8 | zero-shot | Official leaderboard | |
| 15 | chatlaw-33b-hf | 26.1 | zero-shot | Official leaderboard | |
| 16 | tigerbot-sft-7b-hf | 25.9 | zero-shot | Official leaderboard | |
| 17 | lawyer-llama-13b-hf | 25.3 | zero-shot | Official leaderboard | |
| 18 | baichuan-13b-base-hf | 24.0 | one-shot | Official leaderboard | |
| 19 | chatglm2-6b | 22.9 | one-shot | Jun 2023 | Official leaderboard |
| 20 | internlm-7b | 22.5 | one-shot | Jul 2023 | Official leaderboard |
| 21 | tigerbot-base-7b-hf | 21.7 | zero-shot | Official leaderboard | |
| 22 | baichuan-7b-hf | 21.2 | one-shot | Official leaderboard | |
| 23 | | 20.8 | one-shot | Jul 2023 | Official leaderboard |
| 24 | | 19.6 | zero-shot | Sep 2023 | Official leaderboard |
| 25 | chinese-alpaca-2-7b-hf | 19.2 | one-shot | Official leaderboard | |
| 26 | vicuna-33b-hf | 19.1 | one-shot | Official leaderboard | |
| 27 | ziya-llama-13b-hf | 18.8 | one-shot | Official leaderboard | |
| 28 | | 18.4 | one-shot | Jul 2023 | Official leaderboard |
| 29 | | 18.3 | one-shot | Jul 2023 | Official leaderboard |
| 30 | | 17.7 | one-shot | Feb 2023 | Official leaderboard |
| 31 | llama-2-chinese-13b-hf | 17.1 | one-shot | Official leaderboard | |
| 32 | | 16.6 | one-shot | Feb 2023 | Official leaderboard |
| 33 | llama-30b | 16.5 | one-shot | Official leaderboard | |
| 34 | gogpt-7b | 16.2 | one-shot | Official leaderboard | |
| 35 | vicuna-13b-hf | 15.8 | one-shot | Official leaderboard | |
| 36 | moss-moon-003-base-hf | 15.7 | one-shot | Official leaderboard | |
| 37 | alpaca-7b-hf | 15.6 | one-shot | Official leaderboard | |
| 38 | moss-moon-003-sft-hf | 15.6 | one-shot | Official leaderboard | |
| 39 | | 14.8 | one-shot | Feb 2023 | Official leaderboard |
| 40 | llama-2-70b-chat | 14.7 | zero-shot | Official leaderboard | |
| 41 | llama-2-13b-chat | 14.6 | zero-shot | Official leaderboard | |
| 42 | wizardlm-7b-hf | 14.5 | one-shot | Official leaderboard | |
| 43 | | 14.3 | one-shot | May 2023 | Official leaderboard |
| 44 | xverse-13b-hf | 13.7 | one-shot | Official leaderboard | |
| 45 | vicuna-7b-hf | 13.7 | one-shot | Official leaderboard | |
| 46 | mpt-instruct-7b-hf | 12.6 | one-shot | Official leaderboard | |
| 47 | chinese-llama-2-7b-hf | 12.6 | one-shot | Official leaderboard | |
| 48 | lawgpt-7b-beta1.1-hf | 11.6 | one-shot | Official leaderboard | |
| 49 | llama-2-chinese-7b-hf | 10.7 | one-shot | Official leaderboard | |
| 50 | llama-2-7b-chat | 7.16 | zero-shot | Official leaderboard | |
| 51 | lawgpt-7b-beta1.0-hf | 2.33 | one-shot | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-3.5 Turbo Jun 2023 | 44.5 | one-shot | Jun 2023 | Official leaderboard |
| 2 | chatglm2-6b | 22.9 | one-shot | Jun 2023 | Official leaderboard |
| 3 | internlm-7b | 22.5 | one-shot | Jul 2023 | Official leaderboard |
| 4 | | 20.8 | one-shot | Jul 2023 | Official leaderboard |
| 5 | | 19.6 | zero-shot | Sep 2023 | Official leaderboard |
| 6 | | 18.4 | one-shot | Jul 2023 | Official leaderboard |
| 7 | | 18.3 | one-shot | Jul 2023 | Official leaderboard |
| 8 | | 17.7 | one-shot | Feb 2023 | Official leaderboard |
| 9 | | 16.6 | one-shot | Feb 2023 | Official leaderboard |
| 10 | | 14.8 | one-shot | Feb 2023 | Official leaderboard |
| 11 | | 14.3 | one-shot | May 2023 | Official leaderboard |
Rankings by part
We read 20 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Task 1-1, Task 1-2 and Task 2-1. GPT-3.5 Turbo leads 16 of them, and 2 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.
Task 1-1
11 models we track have a score here.
- 1 GPT-3.5 Turbo Jun 2023 16.1
- 2 chatglm2-6b 15.4
- 3
Llama 2-13B ↑3 3.58
- 4
MPT-7B ↑7 3.57
- 5
Llama 2-7B ↑2 3.42
- 6
Llama 2-70B ↓2 3.33
- 7
LLaMA-7B ↑3 3.14
- 8
LLaMA-13B ↑1 3.00
- 9 internlm-7b ↓6 2.79
- 10
LLaMA-65B ↓2 2.00
Task 1-2
11 models we track have a score here.
- 1 GPT-3.5 Turbo Jun 2023 37.2
- 2 chatglm2-6b 24.8
- 3 internlm-7b 4.60
- 4
LLaMA-13B ↑5 2.60
- 5
Llama 2-13B ↑1 1.80
- 6
Llama 2-7B ↑1 1.80
- 7
LLaMA-65B ↑1 1.20
- 8
LLaMA-7B ↑2 1.20
- 9
MPT-7B ↑2 0.40
- 10
Llama 2-70B ↓6 0.40
Task 2-1
11 models we track have a score here.
- 1 GPT-3.5 Turbo Jun 2023 13.5
- 2 internlm-7b ↑1 10.7
- 3
Llama 2-70B ↑1 9.93
- 4
Qwen-7B ↑1 8.12
- 5
LLaMA-65B ↑3 3.94
- 6
Llama 2-13B 3.73
- 7 chatglm2-6b ↓5 3.17
- 8
MPT-7B ↑3 2.35
- 9
Llama 2-7B ↓2 1.42
- 10
LLaMA-7B 1.16
Task 2-2
11 models we track have a score here.
- 1 GPT-3.5 Turbo Jun 2023 40.6
- 2
Llama 2-70B ↑2 11.0
- 3 chatglm2-6b ↓1 10.8
- 4
LLaMA-65B ↑4 9.20
- 5
Qwen-7B 8.80
- 6
Llama 2-13B 8.60
- 7
MPT-7B ↑4 7.42
- 8
Llama 2-7B ↓1 6.80
- 9 internlm-7b ↓6 6.40
- 10
LLaMA-13B ↓1 6.39
Task 2-3
11 models we track have a score here.
- 1 GPT-3.5 Turbo Jun 2023 54.0
- 2
Llama 2-70B ↑2 34.1
- 3
Llama 2-13B ↑3 31.5
- 4 internlm-7b ↓1 26.7
- 5
LLaMA-7B ↑5 25.5
- 6
LLaMA-65B ↑2 23.9
- 7
Qwen-7B ↓2 23.2
- 8
LLaMA-13B ↑1 22.5
- 9
MPT-7B ↑2 22.4
- 10
Llama 2-7B ↓3 20.6
Task 2-4
11 models we track have a score here.
- 1 GPT-3.5 Turbo Jun 2023 41.4
- 2
Qwen-7B ↑3 24.8
- 3 chatglm2-6b ↓1 22.4
- 4
Llama 2-70B 10.4
- 5 internlm-7b ↓2 8.40
- 6
LLaMA-7B ↑4 7.40
- 7
MPT-7B ↑4 6.80
- 8
Llama 2-7B ↓1 6.20
- 9
Llama 2-13B ↓3 2.80
- 10
LLaMA-13B ↓1 2.40
Task 2-5
11 models we track have a score here.
- 1 GPT-3.5 Turbo Jun 2023 62.0
- 2 chatglm2-6b 38.4
- 3
Qwen-7B ↑2 31.7
- 4
Llama 2-70B 27.3
- 5
LLaMA-65B ↑3 25.3
- 6
Llama 2-13B 24.9
- 7
Llama 2-7B 23.5
- 8
LLaMA-7B ↑2 21.9
- 9
LLaMA-13B 20.5
- 10 internlm-7b ↓7 6.04
Task 2-6
11 models we track have a score here.
- 1 GPT-3.5 Turbo Jun 2023 74.0
- 2
Llama 2-70B ↑2 42.6
- 3
Qwen-7B ↑2 42.5
- 4 chatglm2-6b ↓2 36.8
- 5 internlm-7b ↓2 35.6
- 6
MPT-7B ↑5 34.9
- 7
Llama 2-7B 34.5
- 8
LLaMA-13B ↑1 30.9
- 9
Llama 2-13B ↓3 25.8
- 10
LLaMA-65B ↓2 20.3
Task 2-7
11 models we track have a score here.
- 1 GPT-3.5 Turbo Jun 2023 40.7
- 2
Llama 2-7B ↑5 31.0
- 3
LLaMA-7B ↑7 21.7
- 4
Llama 2-13B ↑2 20.5
- 5
LLaMA-13B ↑4 19.7
- 6 chatglm2-6b ↓4 19.5
- 7
Llama 2-70B ↓3 16.5
- 8
MPT-7B ↑3 16.1
- 9 internlm-7b ↓6 14.6
- 10
LLaMA-65B ↓2 13.7
Task 2-8
11 models we track have a score here.
- 1 GPT-3.5 Turbo Jun 2023 37.4
- 2 chatglm2-6b 12.8
- 3
Qwen-7B ↑2 7.40
- 4
Llama 2-70B 6.20
- 5
LLaMA-13B ↑4 5.00
- 6
LLaMA-65B ↑2 4.60
- 7
Llama 2-7B 4.60
- 8 internlm-7b ↓5 2.80
- 9
MPT-7B ↑2 2.40
- 10
LLaMA-7B 1.60
Task 2-9
11 models we track have a score here.
- 1 GPT-3.5 Turbo Jun 2023 67.6
- 2 internlm-7b ↑1 35.0
- 3
Qwen-7B ↑2 34.9
- 4
Llama 2-70B 34.3
- 5
LLaMA-65B ↑3 33.7
- 6
Llama 2-13B 31.0
- 7 chatglm2-6b ↓5 26.7
- 8
Llama 2-7B ↓1 25.3
- 9
LLaMA-13B 18.1
- 10
LLaMA-7B 14.4
Task 2-10
11 models we track have a score here.
- 1 GPT-3.5 Turbo Jun 2023 40.0
- 2
Qwen-7B ↑3 8.39
- 3 chatglm2-6b ↓1 5.64
- 4
Llama 2-70B 3.84
- 5
Llama 2-13B ↑1 3.08
- 6
LLaMA-65B ↑2 2.17
- 7 internlm-7b ↓4 1.58
- 8
MPT-7B ↑3 1.53
- 9
LLaMA-13B 1.39
- 10
Llama 2-7B ↓3 1.03
Task 3-1
11 models we track have a score here.
- 1 internlm-7b ↑2 42.5
- 2 GPT-3.5 Turbo Jun 2023 ↓1 30.8
- 3 chatglm2-6b ↓1 23.1
- 4
Qwen-7B ↑1 19.9
- 5
Llama 2-70B ↓1 19.5
- 6
LLaMA-65B ↑2 6.00
- 7
Llama 2-13B ↓1 3.60
- 8
Llama 2-7B ↓1 1.29
- 9
LLaMA-13B 1.01
- 10
LLaMA-7B 0.95
Task 3-2
11 models we track have a score here.
- 1 GPT-3.5 Turbo Jun 2023 34.5
- 2 chatglm2-6b 34.2
- 3
Qwen-7B ↑2 20.8
- 4
Llama 2-13B ↑2 12.0
- 5
Llama 2-70B ↓1 11.9
- 6
Llama 2-7B ↑1 10.4
- 7
LLaMA-65B ↑1 10.3
- 8
LLaMA-7B ↑2 10.2
- 9
LLaMA-13B 10.00
- 10
MPT-7B ↑1 9.82
Task 3-3
11 models we track have a score here.
- 1 internlm-7b ↑2 41.4
- 2 GPT-3.5 Turbo Jun 2023 ↓1 35.5
- 3 chatglm2-6b ↓1 32.1
- 4
Qwen-7B ↑1 30.6
- 5
Llama 2-70B ↓1 18.5
- 6
Llama 2-13B 15.1
- 7
Llama 2-7B 12.5
- 8
LLaMA-65B 12.5
- 9
LLaMA-13B 12.4
- 10
LLaMA-7B 10.4
Task 3-4
11 models we track have a score here.
- 1 GPT-3.5 Turbo Jun 2023 78.7
- 2
Qwen-7B ↑3 77.9
- 3 internlm-7b 76.7
- 4
Llama 2-7B ↑3 73.1
- 5
LLaMA-13B ↑4 72.4
- 6
LLaMA-65B ↑2 71.8
- 7
Llama 2-13B ↓1 68.3
- 8
Llama 2-70B ↓4 66.1
- 9
MPT-7B ↑2 62.2
- 10 chatglm2-6b ↓8 59.8
Task 3-5
11 models we track have a score here.
- 1 internlm-7b ↑2 77.1
- 2 GPT-3.5 Turbo Jun 2023 ↓1 76.8
- 3 chatglm2-6b ↓1 72.8
- 4
LLaMA-65B ↑4 60.9
- 5
Llama 2-70B ↓1 60.4
- 6
Llama 2-7B ↑1 57.4
- 7
Qwen-7B ↓2 56.8
- 8
MPT-7B ↑3 56.1
- 9
Llama 2-13B ↓3 55.0
- 10
LLaMA-7B 54.2
Task 3-6
11 models we track have a score here.
- 1 GPT-3.5 Turbo Jun 2023 31.6
- 2 chatglm2-6b 16.6
- 3 internlm-7b 10.4
- 4
Llama 2-7B ↑3 8.80
- 5
LLaMA-65B ↑3 6.80
- 6
LLaMA-7B ↑4 6.80
- 7
Llama 2-13B ↓1 6.60
- 8
LLaMA-13B ↑1 6.40
- 9
Llama 2-70B ↓5 2.80
- 10
MPT-7B ↑1 2.00
Task 3-7
11 models we track have a score here.
- 1 GPT-3.5 Turbo Jun 2023 66.4
- 2
Llama 2-70B ↑2 53.0
- 3
Qwen-7B ↑2 44.6
- 4 chatglm2-6b ↓2 42.8
- 5 internlm-7b ↓2 38.0
- 6
Llama 2-13B 37.4
- 7
LLaMA-13B ↑2 33.6
- 8
LLaMA-65B 32.8
- 9
MPT-7B ↑2 32.8
- 10
Llama 2-7B ↓3 32.0
Task 3-8
11 models we track have a score here.
- 1 chatglm2-6b ↑1 17.6
- 2 GPT-3.5 Turbo Jun 2023 ↓1 17.4
- 3
Llama 2-70B ↑1 13.4
- 4
Llama 2-13B ↑2 12.2
- 5 internlm-7b ↓2 11.3
- 6
Llama 2-7B ↑1 11.0
- 7
LLaMA-65B ↑1 10.5
- 8
LLaMA-7B ↑2 10.5
- 9
LLaMA-13B 10.1
- 10
Qwen-7B ↓5 10.1
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: Official LawBench leaderboard
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| GPT-3.5 Turbo Jun 2023 | one-shot | 44.5 | |||
| GPT-3.5 Turbo Jun 2023 | zero-shot | 42.1 | |||
| chatglm2-6b | one-shot | 22.9 | |||
| internlm-7b | one-shot | 22.5 | |||
| chatglm2-6b | zero-shot | 21.1 | |||
| one-shot | 20.8 | ||||
| zero-shot | 19.6 | ||||
| one-shot | 18.6 | ||||
| one-shot | 18.4 | ||||
| one-shot | 18.3 | ||||
| one-shot | 17.7 | ||||
| one-shot | 16.6 | ||||
| internlm-7b | zero-shot | 16.3 | |||
| one-shot | 14.8 | ||||
| one-shot | 14.3 | ||||
| zero-shot | 13.6 | ||||
| zero-shot | 11.7 | ||||
| zero-shot | 11.2 | ||||
| zero-shot | 9.78 | ||||
| zero-shot | 9.72 | ||||
| zero-shot | 8.74 | ||||
| zero-shot | 8.44 |
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: Official LawBench leaderboard
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| GPT-3.5 Turbo Jun 2023 | one-shot | 44.5 | |||
| GPT-3.5 Turbo Jun 2023 | zero-shot | 42.1 | |||
| chatglm2-6b | one-shot | 22.9 | |||
| internlm-7b | one-shot | 22.5 | |||
| chatglm2-6b | zero-shot | 21.1 | |||
| one-shot | 20.8 | ||||
| zero-shot | 19.6 | ||||
| one-shot | 18.6 | ||||
| one-shot | 18.4 | ||||
| one-shot | 18.3 | ||||
| one-shot | 17.7 | ||||
| one-shot | 16.6 | ||||
| internlm-7b | zero-shot | 16.3 | |||
| one-shot | 14.8 | ||||
| one-shot | 14.3 | ||||
| zero-shot | 13.6 | ||||
| zero-shot | 11.7 | ||||
| zero-shot | 11.2 | ||||
| zero-shot | 9.78 | ||||
| zero-shot | 9.72 | ||||
| zero-shot | 8.74 | ||||
| zero-shot | 8.44 |
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official LawBench leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official LawBench 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
8 of 9 checks pass, 1 warning. Look first at the Newest score check. The newest came 3 years ago, in September 2023, for Qwen-7B. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 1 warning 2 pass 2 notes
- Pass Independent runs Scores from the official leaderboard (51 entries).
- Pass Models scored 11 models we track, plus 40 other entries on its leaderboard.
- Warning Newest score The newest came 3 years ago, in September 2023, for Qwen-7B.
- Note Frontier models None of the top 10 has a score. The strongest tested is GPT-3.5 Turbo (#209).
- Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
- Pass Importer We read the board on github.com: 22 rows for models we track and 40 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.