HELM Lite - LegalBench - International Citizenship Questions: leaderboard
HELM Lite v1.13.0 LegalBench International Citizenship Questions task: answer yes/no questions about how national citizenship laws govern acquiring and losing citizenship, drawn from the GLOBALCIT database; 1,000 items per model, exact match. Percent correct, 0-100, higher is better.
Metric: Exact Match (%). Source: crfm.stanford.edu. Saturation forecast: Estimated already saturated. 91 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Llama 3 70B | 70.5 | #632 |
| 2 | Gemini 1.5 Pro (002) | 68.5 | #368 |
| 3 | Gemini 1.5 Pro (001) | 67.5 | #465 |
| 4 | Nova Lite | 65.8 | #741 |
| 5 | GPT-4 Turbo | 65.6 | #417 |
| 6 | Nova Pro | 65.5 | #594 |
| 7 | Qwen 2 72B Instruct | 65.3 | #579 |
| 8 | Qwen 1.5 7B | 63.3 | #1297 |
| 9 | Qwen 2.5 72B Instruct | 63.1 | #469 |
| 10 | GPT-4 (0613) | 63 | #493 |
| 11 | Nova Micro | 62.9 | #891 |
| 12 | Llama 2 70B | 62 | #1039 |
| 13 | Llama 2 13B | 61.8 | #1411 |
| 14 | Qwen 1.5 72B | 61.2 | #752 |
| 15 | Qwen 1.5 14B | 60.9 | #1083 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=helm-lite-legalbench-international-citizenship-questions · How It Works · Data refreshed daily, snapshot 2026-11-05.