Atlas

Benchmarks

← All benchmarks

Harvey HLAB - Criteria Pass Rate - Banking Finance

Professional Work · 2026-05-06

The Banking Finance 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
Muse Spark 1.297.8
Muse Spark 1.397.4
Gemini 4 Argon95.9
Grok 4.595.5
Grok 4.694.8
Muse Spark 1.194.8
Gemini 3.8 Flash94.4
Muse Wonder94.4
GLM-5.392.6
Muse Wonder Full92.6
GLM-5.3 Flash92.2
Grok 4.792.2
Sonnet 4.691.5
Mistral Large 4 Preview91.5
Qwen3.8-Max91.5
Opus 4.791.1
Gemini 3.7 Flash91.1
MiMo-V2.6-Flash91.1
MiMo-V2.6-Pro91.1
GPT-6.1 Sol90.7
Qwen3.8 27B90.7
Kimi K390.3
Sonnet 590.0
MiMo-V2.590.0
Claude Sonnet 5.589.6
DeepSeek V4.1 Flash 101789.6
DeepSeek V4 Pro 081389.6
Step 5 Preview89.6
Hy4 Preview89.2
Claude Opus 5.588.8
Ember-188.1
GPT-6 Astra88.1
Opus 4.887.7
GPT-6 Sol87.7
DeepSeek V4.1 Flash87.4
DeepSeek-V4-Flash-073187.4
GPT-5.6 Luna86.6
MiniMax M386.6
Claude Opus 586.2
DeepSeek-V4-Pro86.2
Loading Atlas data…