Harvey HLAB — Criteria Pass Rate
Professional Work · 2026-05-06
Harvey's Legal Agent Benchmark held-out evaluation criteria pass rate (0–100). Split from the coarse task score, which is on a different scale.
Top models (higher is better)
| Model | Score |
|---|---|
| Muse Spark 1.1 | 92.9 |
| Kimi K3 | 90.8 |
| Grok 4.5 | 90.5 |
| Claude Fable 5 | 90.5 |
| Opus 4.8 | 87.9 |
| Claude Opus 5 | 87.7 |
| Sonnet 4.6 | 86.7 |
| Sonnet 5 | 86.3 |
| MiniMax M3 | 86.3 |
| GLM-5.2 | 85.6 |
| DeepSeek-V4-Pro | 84.4 |
| Gemini 3.6 Flash | 84.4 |
| Qwen3.7-Max | 83.5 |
| GPT-5.6 Luna | 83.4 |
| GPT-5.6 Sol | 80.5 |
| GPT-5.5 | 80.0 |
| Gemini 3.5 Flash | 79.3 |
| GPT-5.6 Terra | 78.5 |
| GPT-5.4 | 74.5 |
| Kimi K2.6 | 74.1 |
| GLM-5.1 | 71.7 |
| Gemini 3.5 Flash-Lite | 69.5 |
| Inkling | 65.7 |
| Grok 4.3 | 64.8 |
| Qwen3.7-Plus | 62.7 |
| Laguna M.1 | 54.9 |
| Laguna XS.2 | 41.1 |