Vals AI Harvey Legal Agent Bench: leaderboard
Harvey's Legal Agent Benchmark (Vals AI) — agentic legal work across documents, spreadsheets, presentations, and file-system tools spanning practice areas (M&A, antitrust, capital markets, tax); final score percentage.
Metric: Accuracy (%). Source: www.vals.ai. Saturation forecast: Around 2029. 81 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Muse Spark 1.2 | 25.42 | #62 |
| 2 | Muse Spark 1.3 (Max) | 23.75 | #24 (Muse Spark 1.3) |
| 3 | Muse Spark 1.3 | 22.92 | #24 |
| 4 | Muse Wonder Full | 21.67 | #10 |
| 5 | Muse Spark 1.1 | 20 | #52 |
| 6 | Gemini 4 Argon | 19.58 | #8 |
| 7 | Muse Wonder | 17.08 | #15 |
| 8 | Muse Spark Open | 16.67 | #54 |
| 9 | Grok 4.6 | 15.83 | #26 |
| 10 | Mistral Large 4 | 15.83 | #94 |
| 11 | Kimi K3 | 12.92 | #30 |
| 12 | Grok 4.5 | 12.92 | #48 |
| 13 | Grok 4.7 | 12.5 | #25 |
| 14 | Qwen 3.8 27B | 11.25 | #105 |
| 15 | Claude Fable 5 | 11.25 | #16 |
Interactive version: theaggregate.ai/benchmark?slug=vals-ai-harvey-legal-agent-bench · How It Works · Data refreshed daily, snapshot 2026-10-20.