Share your thoughts, 1 month free Claude Pro on usSee more

Legal Evaluation on LawBench

62Accuracy

LawThinker

Updated 1mo ago

Evaluation Results

Method	Links
LawThinker 2026.02		62	-	-	-
PoliLegalLM-30B-A3B 2026.04		60.95	68.56	54	67.73
Qwen3-235B-A22B 2026.04		59.74	66.72	52.38	67.2
GPT-5 2026.04		59.16	45.68	56.89	65.38
GLM-5 2026.04		58.74	67.47	50.33	67.07
Plan-and-Solve 2026.02		57.9	-	-	-
Plan-and-Execute 2026.02		57	-	-	-
PoliLegalLM-32B 2026.04		56.26	48.82	50.85	64.87
ReAct 2026.02		56.2	-	-	-
Search-o1 2026.02		55.3	-	-	-
DeepSeek-V3.2 2026.04		54.43	64.32	44.76	64.05
GPT-4 2026.06		53.85	-	-	-
Grok-4.2 2026.04		53.33	51.66	45.43	63.63
Doubao-Seed-2.0-Pro 2026.04		52.65	45.06	53.63	53.32
Direct Reasoning 2026.02		52	-	-	-
Qwen3.5-397B-A17B 2026.04		50.95	70.69	50.74	46.27
Kimi-K2.5 2026.04		48.59	33.47	43.66	58.52
Qwen3-4B-J 2026.06		48.28	-	-	-
Qwen3-8B-J 2026.06		47.99	-	-	-
Qwen3-8B 2026.06		47.22	-	-	-
Qwen3-4B 2026.06		47.08	-	-	-
Qwen3-4B-PD 2026.06		46.94	-	-	-
Qwen3-8B-PD 2026.06		46.92	-	-	-
Qwen-Chat-7B 2026.06		38.99	-	-	-
ChatLaw-33B 2026.06		29.14	-	-	-
Fuzi-Mingcha 2026.06		28.78	-	-	-