Atlas

Benchmarks

← All benchmarks

Harvey HLAB - Criteria Pass Rate - Bankruptcy Restructuring

Professional Work · 2026-05-06

The Bankruptcy Restructuring 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 1.395.5
Muse Spark 1.293.8
Grok 4.692.0
Kimi K392.0
Muse Spark 1.192.0
Gemini 4 Argon91.3
Muse Wonder Full91.0
MiMo-V2.6-Pro90.3
Qwen3.8-Max90.3
Muse Wonder90.0
Mistral Large 4 Preview89.6
Ember-189.3
Opus 4.788.2
Grok 4.588.2
DeepSeek-V4-Flash-073187.9
GLM-5.387.5
Gemini 3.8 Flash87.2
Grok 4.787.2
MiMo-V2.6-Flash87.2
DeepSeek V4 Pro 081386.9
Sonnet 4.685.8
Qwen3.8 27B85.8
DeepSeek V4.1 Flash 101785.5
GPT-6.1 Sol85.5
Hy4 Preview85.5
Step 5 Preview85.5
Claude Opus 584.1
Gemini 3.7 Flash84.1
Sonnet 583.7
Opus 4.883.0
GPT-6 Astra82.0
GPT-5.6 Luna81.3
GPT-6 Sol81.3
GPT-6 Luna80.6
DeepSeek V4.1 Flash80.3
Gemini 3.6 Flash80.3
MiniMax M380.3
Qwen3.7-Max (source unspecified)78.2
Claude Haiku 5.577.9
Muse Spark Open76.8
Loading Atlas data…