Atlas

Benchmarks

← All benchmarks

Harvey HLAB - Criteria Pass Rate - Data Privacy Cybersecurity

Professional Work · 2026-05-06

The Data Privacy Cybersecurity practice-area split of Harvey's Legal Agent Benchmark held-out evaluation criteria pass rate. This reports pooled rubric criteria passed within that split and is distinct from the all-pass task-score split.

Top models (higher is better)

ModelScore
GLM-5.295.1
Grok 4.595.1
Claude Fable 594.3
Opus 4.894.3
Claude Opus 594.3
Kimi K394.0
Muse Spark 1.193.7
GPT-5.592.2
MiniMax M392.2
Gemini 3.6 Flash91.5
GPT-5.6 Luna91.5
Sonnet 591.2
GPT-5.6 Sol90.1
Sonnet 4.689.0
GLM-5.188.7
Qwen3.7-Max88.3
GPT-5.6 Terra86.2
DeepSeek-V4-Pro85.9
Gemini 3.5 Flash84.5
GPT-5.479.2
Kimi K2.674.6
Grok 4.373.9
Inkling72.8
Gemini 3.5 Flash-Lite72.4
Qwen3.7-Plus64.0
Laguna M.151.9
Laguna XS.246.7
Loading Atlas data…