Benchmarks / Professional Domains
CaseLaw
Vals AI's case-law research and citation benchmark.
maintained by Vals AI · www.vals.ai · 54 observations on record
Versions
1 version
| Version | Tasks |
|---|---|
| CaseLaw current case-law | — |
Current leaderboard
CaseLaw · Accuracy · top 25 of 54 systems
| # | System | Accuracy % |
|---|---|---|
| 1 | Grok 4.3 xAI | 79.3% |
| 2 | GPT 5.1 (2025-11-13)high effort OpenAI | 73.4% |
| 3 | GPT 4.1 (2025-04-14)high effort OpenAI | 69.9% |
| 4 | GPT 5 Mini (2025-08-07)high effort OpenAI | 68.5% |
| 5 | Claude Opus 4.7 Anthropic | 68.4% |
| 6 | GPT 5 (2025-08-07)high effort OpenAI | 66.5% |
| 7 | GPT 5.5 (2026-04-22)xhigh effort OpenAI | 66.2% |
| 8 | GPT 5.2 (2025-12-11)xhigh effort OpenAI | 66.0% |
| 9 | Grok 4 0709 xAI | 65.8% |
| 10 | Grok 4 Fast Reasoning xAI | 65.7% |
| 11 | Kimi K2thinking Moonshot AI | 65.7% |
| 12 | Gemini 3.1 Pro Previewhigh effort Google DeepMind | 64.8% |
| 13 | Command A 03 2025 cohere | 64.5% |
| 14 | Claude Sonnet 4.6 Anthropic | 64.0% |
| 15 | Gemini 2.5 Pro (2025-06-17) Google DeepMind | 63.9% |
| 16 | GPT 5.4 (2026-03-05)xhigh effort OpenAI | 63.8% |
| 17 | Muse Spark Meta AI | 63.1% |
| 18 | Claude Opus 4.5 20251101thinking Anthropic | 62.6% |
| 19 | Claude Sonnet 4.5 20250929thinking Anthropic | 62.2% |
| 20 | Claude Opus 4.6thinking Anthropic | 62.1% |
| 21 | Mistral Large 2512 Mistral AI | 61.4% |
| 22 | Kimi K2 6 Moonshot AI | 61.2% |
| 23 | Minimax M2 7 MiniMax | 60.9% |
| 24 | Grok 4.1 Fast Reasoning xAI | 60.5% |
| 25 | GPT 4o (2024-11-20)high effort OpenAI | 59.7% |
One row per evaluated system — reasoning-effort variants rank separately.