HELM EWoK - Quantitative Properties: leaderboard
Metric: Exact Match (%). Source: crfm.stanford.edu. 15 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 2 72B Instruct | 100 |
| 2 | Claude 3 Opus (20240229) | 100 |
| 3 | Claude 3.5 Sonnet (20240620) | 99.47 |
| 4 | GPT-4 Turbo | 98.68 |
| 5 | Claude 3 Sonnet (20240229) | 98.42 |
| 6 | Llama 3 70B Instruct | 98.16 |
| 7 | GPT-4o (2024-05-13) | 98.16 |
| 8 | GPT-4o Mini (2024-07-18) | 97.63 |
| 9 | Claude 3 Haiku (20240307) | 96.84 |
| 10 | Mistral 7B Instruct (v0.3) | 91.32 |
| 11 | GPT-3.5 Turbo (0125) | 90.79 |
| 12 | Claude 3.5 Haiku (20241022) | 90.53 |
| 13 | Llama 3 8B Instruct | 17.63 |
Interactive version: theaggregate.ai/benchmark?slug=helm-ewok-quantitative-properties · How It Works · Data refreshed daily, snapshot 2026-09-19.