Harvey HLAB - Criteria Pass Rate - Arbitration International Dispute Resolution
Professional Work · 2026-05-06
The Arbitration International Dispute Resolution 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 | 94.3 |
| Grok 4.5 | 92.1 |
| Kimi K3 | 92.1 |
| GLM-5.2 | 91.8 |
| Claude Fable 5 | 91.4 |
| Claude Opus 5 | 90.3 |
| Opus 4.8 | 88.2 |
| Sonnet 5 | 87.5 |
| MiniMax M3 | 86.7 |
| Sonnet 4.6 | 85.7 |
| DeepSeek-V4-Pro | 83.5 |
| GPT-5.6 Sol | 82.1 |
| GPT-5.6 Luna | 81.7 |
| Gemini 3.6 Flash | 78.9 |
| Qwen3.7-Max | 78.9 |
| Gemini 3.5 Flash | 78.5 |
| GPT-5.5 | 78.5 |
| GPT-5.6 Terra | 74.6 |
| Grok 4.3 | 68.1 |
| GPT-5.4 | 66.7 |
| Gemini 3.5 Flash-Lite | 63.8 |
| Qwen3.7-Plus | 58.4 |
| GLM-5.1 | 53.4 |
| Kimi K2.6 | 50.2 |
| Inkling | 49.5 |
| Laguna XS.2 | 43.7 |
| Laguna M.1 | 41.9 |