HELM Classic - LegalSupport: leaderboard
The "LegalSupport" subtask of HELM Classic. A scenario from Stanford HELM Classic, the holistic evaluation suite scoring language models across accuracy, robustness and calibration on a broad set of tasks.
Metric: Exact Match (%). Source: crfm.stanford.edu. Saturation forecast: Estimated already saturated. 69 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | LLaMA-30B | 63.8 | #1423 |
| 2 | Palmyra X (43B) | 62.3 | #1317 |
| 3 | text-davinci-003 | 62.17 | #1146 |
| 4 | text-davinci-002 | 61.49 | #1225 |
| 5 | vicuna-7B-v1.3 | 60.53 | #1507 |
| 6 | falcon-40B Instruct | 60.53 | #1458 |
| 7 | falcon-40B | 60.53 | #1442 |
| 8 | Llama 2 13B | 59.1 | #1436 |
| 9 | LLaMA-65B | 59.1 | #1303 |
| 10 | vicuna-13B-v1.3 | 58.9 | #1447 |
| 11 | LLaMA-13B | 58.69 | #1516 |
| 12 | Mistral-7B-v0.1 | 58.49 | #1326 |
| 13 | Llama 2 70B | 58.49 | #1201 |
| 14 | RedPajama-INCITE-7B-Instruct | 56.85 | #1566 |
| 15 | mpt-30B | 56.44 | #1476 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=helm-classic-legalsupport · How It Works · Data refreshed daily, snapshot 2026-10-20.