Atlas

Benchmarks

← All benchmarks

Harvey's Legal Agent Benchmark - Capital Markets

Professional Work · 2026-05-06

The Capital Markets split of Vals' archived v1 Harvey Legal Agent Benchmark held-out evaluation. Its task score averages two judges' all-criteria-pass decisions rather than counting binary task successes.

Top models (higher is better)

ModelScore
Muse Spark 1.260.0
Gemini 4 Argon40.0
Grok 4.540.0
Grok 4.740.0
Muse Spark 1.340.0
Haiku 4.520.0
Opus 4.720.0
Opus 4.820.0
Sonnet 4.620.0
DeepSeek-V4-Pro20.0
Gemini 3.6 Flash20.0
Gemini 3.7 Flash20.0
Gemini 3.8 Flash20.0
Inkling20.0
Inkling-Small20.0
Kimi K320.0
MiniMax M320.0
Muse Spark 1.120.0
Muse Spark Open20.0
Muse Wonder Full20.0
Muse Wonder20.0
Qwen3.8 27B20.0
Step 5 Preview20.0
Claude Haiku 5.50.0
Claude Opus 5.50.0
Claude Opus 50.0
Claude Sonnet 5.50.0
Sonnet 50.0
DeepSeek V4.1 Flash 10170.0
DeepSeek V4.1 Flash0.0
DeepSeek-V4-Flash-07310.0
DeepSeek V4 Pro 08130.0
Ember-10.0
Gemini 3.1 Flash-Lite Preview0.0
Gemini 3.1 Pro Preview0.0
Gemini 3.5 Flash0.0
Gemini 3.5 Flash-Lite0.0
Gemini 3 Flash Preview0.0
GLM-5.10.0
GLM-5.20.0
Loading Atlas data…