Legal Research Bench leaderboard

Agentic US legal research: finding and applying the authority that answers a question. Muse Spark 1.3 Max leads with 55.29%.

Models
82
Leader
Muse Spark 1.3 Max
Source
Vals AI
As of
Oct 24, 2026
  1. 1Muse Spark 1.3 Max55.3%
  2. 2Claude Opus 555.3%
  3. 3Claude Fable 5.155.3%
  4. 4Gemini 4 Argon54.8%
  5. 5Muse Wonder Full53.4%
  6. 6Muse Wonder50.5%
  7. 7Claude Opus 5.550.5%
  8. 8Claude Fable 549.5%
  9. 9GLM-5.349.0%
  10. 10Grok 4.648.1%
Legal Research Bench, accuracy against cost per test
Open weightsProprietaryBest for the price
0%15%30%45%60%$.01$.1$1$10Cost per test, log scale →Muse Spark 1.3 MaxMiMo-V2.6-ProGLM LegalLing 3.0 Flash FinGemini 3.1 Flash-Lite Preview

Full leaderboard

RankModelLabAccuracyCost / testHarnessPrice in / out
1Muse Spark 1.3 MaxMeta55.29%$0.59——
2Claude Opus 5Anthropic55.29%$6.76—$5 / $25
3Claude Fable 5.1Anthropic55.29%$23.06—$10 / $50
4Gemini 4 ArgonGoogle54.81%$6.45—$2 / $10
5Muse Wonder FullMeta53.37%$3.60——
6Muse WonderMeta50.48%$3.98—$4 / $20
7Claude Opus 5.5Anthropic50.48%$25.16—$4 / $20
8Claude Fable 5Anthropic49.52%$9.79—$10 / $50
9GLM-5.3Z.ai49.04%$2.24—$0.079 / $1.40
10Grok 4.6xAI48.08%$1.53—$2 / $6
11GPT-5.6 SolOpenAI48.08%$21.61—$4 / $20
12Claude Sonnet 5.5Anthropic48.08%$14.83—$2 / $10
14Grok 4.7xAI47.12%$4.92—$2 / $6
15MiMo-V2.6-ProXiaomi47.12%$0.18—$0.43 / $0.87
16Kimi K3Moonshot AI46.15%$3.47—$0.24 / $11
17Qwen3.8-MaxAlibaba46.15%$5.31—$2 / $6
19GLM-5.3-FlashZ.ai45.19%$0.12—$0.15 / $0.50
20Hy4 previewTencent45.19%$0.85—$0.75 / $2.25
21Ember-1Fireworks AI44.71%$3.67—$3 / $15
22Muse Spark 1.2Meta43.75%$0.51—$1.25 / $4.25
23Claude Opus 4.8Anthropic43.75%$2.82—$5 / $25
24Claude Haiku 5.5Anthropic43.27%$1.63—$0.10 / $0.50
25Claude Sonnet 5Anthropic41.83%$2.72—$2 / $10
26GLM LegalZ.ai41.83%$0.05——
27DeepSeek-V4.1-Flash-20260910DeepSeek41.35%$0.25—$0.15 / $0.60
28GPT-5.6 TerraOpenAI41.35%$7.91—$2 / $12
29Muse Spark 1.3Meta40.87%$0.65—$1.25 / $4.25
31Step 5 PreviewStepFun40.38%$0.84—$1 / $2.70
32GPT-5.5OpenAI40.38%$7.40—$5 / $30
33GPT-6 AstraOpenAI39.42%$10.58—$10 / $50
34Gemini 3.8 FlashGoogle38.94%$1.63—$0.82 / $4.13
35Claude Opus 4.7Anthropic38.46%$3.06—$5 / $25
36Claude Sonnet 4.6Anthropic38.46%$2.30—$3 / $15
37Muse Spark OpenMeta38.46%$1.56—$1.25 / $4.25
38GPT-6.1 SolOpenAI38.46%$3.38—$2 / $10
39Muse Spark 1.1Meta37.98%$0.38—$1.25 / $4.25
40MiMo-V2.6-FlashXiaomi37.98%$0.07—$0.14 / $0.28
41Grok 4.5xAI37.98%$1.09—$2 / $6
42Kimi K3.1 ProMoonshot AI37.50%$7.23—$2 / $10
43GPT-5.6 LunaOpenAI36.54%$0.85—$0.20 / $1.20
44Qwen3.8-27BAlibaba36.06%$2.16—$0.20 / $2.50
45Gemini 3.7 FlashGoogle34.62%$0.76—$0.75 / $3.75
46MiniMax-M3.1-FlashMiniMax34.13%$0.44—$0.15 / $0.60
47Mistral Large 4Mistral AI31.73%$5.33—$1.36 / $4.18
48GLM-5.2Z.ai31.25%$0.89—$0.034 / $7
49Gemini 3.5 FlashGoogle30.77%$1.31—$1.50 / $9
50DeepSeek-V4-FlashDeepSeek30.29%$0.28—$0.042 / $0.18
51GPT-6 LunaOpenAI30.29%$0.44—$0.10 / $0.50
52MiniMax-M3MiniMax29.81%$0.34—$0.30 / $1.20
53Beam 1.0Reflection AI28.85%N/A——
54Mercury 3Inception28.85%$0.07—$0.10 / $0.35
55GPT-6 SolOpenAI28.85%$4.90—$2 / $10
56InklingThinky28.36%$0.49—$1 / $4.05
57GLM-5.1Z.ai27.89%$0.39—$1.40 / $4.40
58Inkling-SmallThinky25.48%$0.14—$0.45 / $1.20
59Qwen3.7-Max-PreviewAlibaba25.48%$1.40—$1.48 / $4.42
60Gemini 3.6 FlashGoogle25.00%$0.47—$0.75 / $3.75
61DeepSeek-V4-ProDeepSeek23.08%$0.68—$0.66 / $1.98
62Gemini 3.1 Pro PreviewGoogle20.67%$1.10—$2 / $12
63Gemini 3 FlashGoogle18.27%$0.14—$0.50 / $3
64Qwen3.7-PlusAlibaba16.35%$0.30—$0.32 / $1.28
65MiMo-V2.5-ProXiaomi15.87%$0.08—$0.30 / $0.61
66Kimi K2.5Moonshot AI15.87%$0.23—$0.50 / $2.50
67Kimi K2.6Moonshot AI15.87%$1.52—$0.65 / $3.41
68Grok 4.3xAI15.38%$0.20—$1.25 / $2.50
69Ling 3.0 Flash Fin—15.38%$0.03——
70Nemotron 3 UltraNVIDIA15.38%N/A—$0.60 / $2.40
71Qwen3.6-PlusAlibaba14.90%$0.70—$0.33 / $1.95
72Gemini 3.5 Flash-LiteGoogle13.94%$0.12—$0.30 / $2.50
73Grok 4.20xAI13.94%$0.26—$1.25 / $2.50
74GPT-5.4 miniOpenAI12.50%$1.86—$0.75 / $4.50
75Claude Haiku 4.5Anthropic10.58%$0.27—$1 / $5
76MiniMax-M2.7MiniMax10.58%$0.22—$0.30 / $1.20
77MiMo-V2.5Xiaomi9.13%$0.03—$0.12 / $0.24
78Mistral Medium 3.5Mistral AI9.13%$1.31—$1.50 / $7.50
79GPT-5.4 nanoOpenAI6.25%$0.14—$0.20 / $1.25
80Mercury 2.5Inception4.33%$0.06—$0.04 / $0.15
81Gemini 3.1 Flash-Lite PreviewGoogle3.37%$0.01—$0.25 / $1.50
82Nemotron 3.5 LightningFireworks AI2.88%$0.00—$0.07 / $0.20
83Laguna M.1Poolside2.40%N/A——
84Laguna XS.2Poolside0.96%N/A——
85Ling 3.0 FlashinclusionAI0.00%$0.04—$0.021 / $0.063

Related leaderboards

Sources

  • Vals AI, as of Oct 24, 2026
  • Provider list prices, as of Oct 24, 2026

Every number on this page comes from these sources; models without a published number show none. Download the data. Cite as “Nexus Explorer, nexus.inc/models”.