HELM Lite - LegalBench - Corporate Lobbying: leaderboard
HELM Lite v1.13.0 LegalBench Corporate Lobbying task: decide from a bill's title and summary and a company's 10-K business description whether the bill is relevant to the company (yes/no); up to 490 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 3.1 70B Instruct | 87.35 | #567 |
| 2 | Gemini 1.5 Pro (001) | 86.12 | #465 |
| 3 | Llama 3.3 70B Instruct | 85.71 | #541 |
| 4 | GPT-4o Mini (2024-07-18) | 85.51 | #568 |
| 5 | Llama 3.2 90B Vision Instruct | 85.51 | #621 |
| 6 | Llama 3.1 405B Instruct | 85.31 | #470 |
| 7 | Qwen 2.5 72B Instruct | 85.1 | #469 |
| 8 | Gemma 2 9B (IT) | 84.49 | #795 |
| 9 | Gemini 1.5 Pro (002) | 84.29 | #368 |
| 10 | GPT-4o (2024-08-06) | 84.29 | #343 |
| 11 | Command-R+ | 84.29 | #869 |
| 12 | DeepSeek V3 | 84.08 | #329 |
| 13 | Mistral Large 2 (Jul) | 84.08 | #490 |
| 14 | Claude 3.5 Haiku (20241022) | 83.47 | #590 |
| 15 | Nova Pro | 83.47 | #594 |
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-corporate-lobbying · How It Works · Data refreshed daily, snapshot 2026-11-05.