MASLegalBench — leaderboard
Benchmark for evaluating multi-agent systems on deductive legal reasoning, with tasks grounded in GDPR-style rule application.
Metric: Best accuracy (self-reported). Source: benchmarklist.com. 5 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Llama 3.1 8B Instruct | 86.21 |
| 2 | GPT-4o Mini | 84.32 |
| 3 | Qwen 2.5 7B Instruct | 75.47 |
| 4 | Qwen 3 8B | 68.74 |
| 5 | DeepSeek V3.1 | 63.05 |
Interactive version: theaggregate.ai/benchmark?slug=maslegalbench · How the rankings work · Data refreshed daily, snapshot 2026-07-22.