Best LLM

What is the best LLM for legal work?

Claude Fable 5.1, Gemini 4 Carbon, Muse Wonder Full and 5 others are practically level at the top in legal tasks.

Among the ten best, Muse Ember 1.0 is the cheapest at $0.20 per million tokens.

The fastest among them is Claude Opus 5.5, at 75 tokens per second.

Ranked models: 81
Tests: 5
Data as of: Nov 10, 2026

Ranking: Legal

Place and value in the area, plus overall index, price and speed. Higher values are better.

Rank	Model	Value	Overall index	Price (3:1)	Speed
1	Claude Fable 5.1Top group	130.6	124.1	$20	59
2	Gemini 4 CarbonTop group	123.8	127.0	n/a	n/a
3	Muse Wonder FullTop group	122.7	118.1	n/a	n/a
4	Claude Fable 5.2Upper group	120.9	130.9	$16	n/a
5	Claude Opus 5.5Upper group	120.6	125.6	$8	75
6	Muse Ember 1.0Upper group	120.3	110.4	$0.20	n/a
7	Gemini 4 ArgonUpper group	119.7	121.4	$4	46
8	Muse WonderUpper group	119.0	113.3	$8	n/a
9	Claude Opus 5Upper group	116.3	117.4	$10	71
10	Inkling 2 SmallUpper group	115.2	n/a	n/a	n/a
11	Claude Sonnet 5.5Middle group	114.3	121.8	$4	103
12	Claude Fable 5Middle group	111.7	112.8	$20	50
13	Qwen 3.8 Max 0902Middle group	111.1	100.8	$2.48	36
14	DeepSeek V4.1 Flash 1017Middle group	110.9	106.7	$0.26	126
15	Grok 4.7Middle group	110.7	103.3	$3	47
16	Muse Spark OpenMiddle group	110.3	97.5	n/a	n/a
17	Kimi K3Middle group	109.4	101.1	$6	61
18	MiMo V2.6 ProMiddle group	109.0	100.4	$0.54	32
19	GPT-6 AstraMiddle group	108.1	120.2	$20	58
20	Hy4Middle group	107.8	n/a	n/a	n/a

Show all 81 ranked models

Rank	Model	Value	Overall index	Price (3:1)	Speed
21	GLM-5.3Middle group	107.6	100.6	$2.15	81
22	Kimi K3.1 ProMiddle group	107.2	112.6	$4	56
23	GPT-5.6 SolMiddle group	106.9	109.1	$8	75
24	GPT-6.1 SolMiddle group	106.8	118.0	$4	44
25	Muse Spark 1.3Middle group	106.0	107.3	$2	102
26	Mistral Large 4Middle group	105.9	100.8	$1.05	n/a
27	Grok 4.6Middle group	105.3	102.1	$3	58
28	MiMo V2.6 FlashMiddle group	104.2	97.3	$0.18	38
29	Step 5 PreviewMiddle group	103.9	97.1	$1.43	65
30	GPT-5.5Middle group	103.3	101.5	$11.25	87
31	Gemini 3.7 FlashMiddle group	102.8	107.9	n/a	72
32	GPT-5.6 TerraMiddle group	102.6	100.8	$4.50	72
33	Claude Opus 4.8Middle group	101.6	99.5	$10	71
34	DeepSeek-V4.1-FlashMiddle group	101.3	99.5	$0.26	118
35	Hy4 previewMiddle group	101.1	n/a	n/a	n/a
36	Gemini 3.8 FlashMiddle group	101.0	106.8	$1.50	81
37	GLM-5.3-FlashMiddle group	100.9	94.3	$0.24	101
38	Grok 4.5Middle group	99.1	94.8	$3	52
39	Claude Opus 4.7Middle group	98.8	95.1	$10	57
40	Claude Sonnet 5Middle group	97.9	92.1	$4	77
41	MiniMax M3.1 FlashMiddle group	96.7	100.0	$0.26	n/a
42	GPT-5.6 LunaMiddle group	96.2	93.8	$0.45	123
43	DeepSeek V4 Pro 0813Middle group	94.7	n/a	n/a	n/a
44	Gemini 3.6 FlashMiddle group	93.5	97.6	$1.50	110
45	Gemini 3.5 FlashMiddle group	93.4	97.8	n/a	142
46	Claude Haiku 5.5Middle group	93.0	102.8	$0.20	164
47	Claude Sonnet 4.6Middle group	92.8	86.7	$6	38
48	GLM-5.2Middle group	91.7	89.0	$2.15	73
49	Qwen 3.8 27BMiddle group	90.4	82.8	$0.66	66
50	GPT-6 LunaMiddle group	89.9	95.8	$0.20	77
51	DeepSeek V4 Flash 0731Middle group	87.7	n/a	n/a	n/a
52	GLM-5.1Middle group	87.6	82.8	$2.15	39
53	Qwen 3.7 MaxMiddle group	87.5	85.4	$2.48	50
54	GPT-5Middle group	86.6	77.5	$3.44	83
55	Gemini 3.1 Pro PreviewLower group	84.9	91.4	$4.50	96
56	Mercury 3Lower group	84.6	n/a	$0.16	n/a
57	GPT-5.4Lower group	84.0	86.5	$5.63	58
58	DeepSeek-V4-ProLower group	83.2	87.2	$0.99	44
59	Gemini 3 Flash PreviewLower group	82.9	81.0	$1.13	101
60	MiniMax M3Lower group	82.6	83.0	$0.52	82
61	Reflection BeamLower group	82.4	82.4	n/a	n/a
62	o3Lower group	82.3	74.2	$3.50	73
63	Kimi K2.6Lower group	80.8	84.9	$1.71	32
64	GPT-6.1 TerraLower group	79.7	106.5	$2	62
65	InklingLower group	76.9	72.1	$1.76	n/a
66	Claude Opus 4.5Lower group	75.0	74.4	$10	44
67	Grok 4.3Lower group	73.8	71.6	$1.56	66
68	Inkling SmallLower group	72.1	68.7	$0.52	61
69	Gemini 3.5 Flash-LiteLower group	71.7	73.7	$0.85	145
70	Nemotron 3 UltraLower group	70.3	66.8	$0.93	92
71	MiMo-V2.5-ProLower group	70.0	73.0	$0.54	59
72	MiniMax M2.7Lower group	67.3	65.1	$0.37	45
73	MiMo-V2.5Lower group	63.4	69.1	$0.18	55
74	Nemotron 3.5 Super VL 120B A12BLower group	62.5	68.1	$0.24	n/a
75	Grok 4.20 ReasoningLower group	61.5	58.3	$1.56	66
76	o4-miniLower group	58.2	63.5	n/a	n/a
77	GPT-5.4 nanoLower group	56.5	64.0	$0.46	98
78	GPT-4.1Lower group	50.1	43.6	$3.50	81
79	Mercury 2.5Lower group	48.8	45.3	n/a	n/a
80	Ling 3.0 FlashLower group	23.2	47.9	$0.032	106
81	Laguna XS.2Lower group	15.9	n/a	n/a	n/a

Price in dollars per million tokens, mixed from 3 parts input and 1 part output. Speed in tokens per second.

The top 20: LegalValue in the area · higher is better

Claude Fable 5.1Rank 1130.6

Gemini 4 CarbonRank 2123.8

Muse Wonder FullRank 3122.7

Claude Fable 5.2Rank 4120.9

Claude Opus 5.5Rank 5120.6

Muse Ember 1.0Rank 6120.3

Gemini 4 ArgonRank 7119.7

Muse WonderRank 8119.0

Claude Opus 5Rank 9116.3

Inkling 2 SmallRank 10115.2

Claude Sonnet 5.5Rank 11114.3

Claude Fable 5Rank 12111.7

Qwen 3.8 Max 0902Rank 13111.1

DeepSeek V4.1 Flash 1017Rank 14110.9

Grok 4.7Rank 15110.7

Muse Spark OpenRank 16110.3

Kimi K3Rank 17109.4

MiMo V2.6 ProRank 18109.0

GPT-6 AstraRank 19108.1

Hy4Rank 20107.8

Which tests count here?

5 tests by independent evaluators feed this area. Each test counts with the result of the model in question.

LegalBench
87 models tested · Vals AI
Legal Research Bench
83 models tested · Vals AI
HLAB-Vals-1.1
35 models tested · Vals AI
Immigration
23 models tested · Vals AI
PRBench Legal
16 models tested · Scale AI

Popular comparisons among the best

Other tasks

All LLMs compared How the overall index works Compare several models