Harvey HLAB - Criteria Pass Rate - Tax
Professional Work · 2026-05-06
The Tax 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)
| Model | Score |
|---|---|
| Muse Spark 1.1 | 91.3 |
| Sonnet 4.6 | 86.0 |
| GLM-5.2 | 85.6 |
| MiniMax M3 | 85.0 |
| Claude Opus 5 | 84.6 |
| Kimi K3 | 84.6 |
| Grok 4.5 | 84.3 |
| Claude Fable 5 | 83.6 |
| Sonnet 5 | 82.3 |
| Opus 4.8 | 80.6 |
| DeepSeek-V4-Pro | 77.9 |
| Gemini 3.6 Flash | 77.6 |
| Qwen3.7-Max | 77.6 |
| Kimi K2.6 | 75.6 |
| GPT-5.6 Luna | 74.9 |
| Gemini 3.5 Flash | 74.6 |
| GPT-5.5 | 72.6 |
| GPT-5.6 Sol | 70.9 |
| GPT-5.6 Terra | 69.9 |
| GLM-5.1 | 67.9 |
| Grok 4.3 | 63.2 |
| Gemini 3.5 Flash-Lite | 62.9 |
| Qwen3.7-Plus | 60.9 |
| GPT-5.4 | 57.5 |
| Laguna M.1 | 56.5 |
| Inkling | 48.8 |
| Laguna XS.2 | 25.6 |