Atlas

Benchmarks

← All benchmarks

Legal Research Bench Weighted Pass Business/Commercial

Professional Work · 2026-06-23

Weighted-pass practice-area score for Business/Commercial tasks in Vals AI Legal Research Bench.

Top models (higher is better)

ModelScore
Claude Opus 592.2
Gemini 4 Argon91.9
GLM-5.3 Flash91.3
Gemini 3.8 Flash91.1
Gemini 4 Carbon91.0
Muse Wonder Full89.8
Grok 4.689.8
Gemini 3.7 Flash89.5
Muse Spark 1.289.2
Muse Spark 1.189.1
Step 5 Preview89.1
GLM-5.389.1
GPT-5.6 Sol89.0
Muse Spark Open89.0
Muse Spark 1.388.7
Opus 4.788.3
Sonnet 4.688.0
DeepSeek V4.1 Flash 101787.8
Hy4 Preview87.7
MiniMax M3.1 Flash87.6
Ember-187.6
GLM-5.287.5
Grok 4.787.4
GPT-5.6 Luna87.3
Claude Haiku 5.587.0
Muse Wonder86.9
Grok 4.586.8
GPT-5.586.8
Sonnet 586.8
Qwen3.8 27B86.8
DeepSeek V4.1 Flash86.7
Qwen3.8-Max86.7
DeepSeek-V4-Pro86.5
GPT-5.6 Terra86.1
Opus 4.886.1
Mistral Large 4 Preview86.0
MiMo-V2.6-Pro85.0
Gemini 3.6 Flash84.6
GPT-6.1 Sol84.2
Gemini 3.5 Flash84.1
Loading Atlas data…