CaseLaw (Vals) · Open Analysis

#	Model	Effort	Score	Cost / task	Harness
1	Grok 4.3	xhigh	79.3%	$0.040	–
2	GPT-5.1	high	73.4%	$0.043	–
3	GPT-4.1	high	69.9%	$0.047	–
4	GPT-5 Mini	high	68.5%	$0.011	–
5	Opus 4.7	max	68.4%	$0.162	–
6	GPT-5	high	66.5%	$0.067	–
7	GPT-5.5	xhigh	66.2%	$0.155	–
8	GPT-5.2	xhigh	66.0%	$0.104	–
9	Grok 4	high	65.8%	$0.159	–
10	Grok 4 Fast	none	65.7%	$0.0073	–
11	Kimi K2	medium	65.7%	$0.015	–
12	Gemini 3.1 Pro	high	64.8%	$0.136	–
13	Command A	none	64.5%	$0.098	–
14	Sonnet 4.6	max	64.0%	$0.145	–
15	Gemini 2.5 Pro	high	63.9%	$0.257	–
16	GPT-5.4	xhigh	63.8%	$0.101	–
17	Muse Spark	–	63.1%	$0.0065	–
18	Opus 4.5	high	62.6%	$0.229	–
19	Sonnet 4.5	high	62.2%	$0.256	–
20	Opus 4.6	max	62.1%	$0.231	–
21	Mistral Large 3	none	61.4%	$0.019	–
22	Kimi K2.6	max	61.2%	$0.026	–
23	MiniMax M2.7	high	60.9%	$0.0031	–
24	Grok 4.1 Fast	high	60.5%	$0.011	–
25	GPT-4o	high	59.7%	$0.067	–
26	Qwen 3.5 Plus	none	59.7%	$0.030	–
27	DeepSeek V4	max	59.4%	$0.013	–
28	Kimi K2.5	max	58.7%	$0.025	–
29	Trinity Large Thinking	–	57.9%	$0.0089	–
30	Haiku 4.5	xhigh	56.5%	$0.050	–
31	Qwen 3.5 Flash	max	55.9%	$0.0049	–
32	MiniMax M2.1	–	55.8%	$0.0080	–
33	Gemini 3 Flash	high	55.8%	$0.025	–
34	DeepSeek V3.2	high	55.4%	$0.018	–
35	Gemini 3.1 Flash-Lite	high	55.0%	$0.0093	–
36	Qwen 3 Max	max	55.0%	$0.085	–
37	GLM 4.7	none	54.9%	$0.021	–
38	Grok 4.20	xhigh	54.4%	$0.068	–
39	DeepSeek V3.1	none	53.9%	$0.017	–
40	MiniMax M2.5	high	53.5%	$0.012	–
41	Qwen 3.6 27B	max	53.2%	$0.024	–
42	Gemini 3 Pro	high	53.1%	$0.138	–
43	Gemma 4 31B IT	high	52.6%	$0.0000	–
44	GPT-5 Nano	high	52.6%	$0.0042	–
45	GLM 5	high	52.5%	$0.034	–
46	GPT-5.4 Nano	high	51.9%	$0.0011	–
47	GPT-5.4 Mini	xhigh	51.7%	$0.041	–
48	GLM 5.1	max	51.6%	$0.034	–
49	Qwen 3.6 Plus	max	51.4%	$0.047	–
50	gpt-oss-120b	high	48.8%	$0.0043	–
51	Qwen 3.6 Max	max	47.9%	$0.052	–
52	Mistral Medium 3.5	high	44.2%	$0.085	–
53	gpt-oss-20b	high	43.8%	$0.0022	–