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

Legal Reasoning on LexEval

63.78Memoization

Kimi-K2-thinking

Updated 1mo ago

Evaluation Results

Method	Links
Kimi-K2-thinking 2026.01		63.78	88.75	76.41	42.9	72.38	68.6	72.81
LegalOne-8B 2026.01		62.68	87.77	74.66	41.63	68.95	67.83	71.19
GPT-5 2026.01		62.31	85.69	73.59	41.07	72.88	73.43	71.78
LegalOne-4B 2026.01		58.59	86.63	72.62	36.87	63.48	60.77	67.59
Qwen3-Max 2026.01		58.46	88	75.82	33	66.88	59.7	68.82
Qwen3-235B-A22B 2026.01		58.32	85.61	73.24	37.47	73.5	62.17	69.47
GLM-4.7 2026.01		56.5	88.31	72.86	33.7	69.18	61.03	68.49
DeepSeek-v3.2 2026.01		55.4	86.67	73.19	29.6	70.2	63.37	68.21
LegalOne-1.7B 2026.01		50.96	82.55	65.63	30.63	55.63	49.93	60.56
DeepSeek-R1 2026.01		49.64	86.91	74.07	33.8	68.88	64.77	68.05
Qwen3-8B 2026.01		49.62	79.31	60.16	29.77	60.65	57.6	60.06
Qwen3-4B 2026.01		44.26	78.49	54.74	23.54	59.15	55.63	56.71
GPT-4o 2026.01		39.74	84.12	65.77	33.97	63.93	58.47	62.32
Qwen3-1.7B 2026.01		29.66	64.29	54.93	24.34	51.5	48.23	49.54
GPT-4o-mini 2026.01		24.19	83.59	50.23	30.9	59.03	42.4	52.91