HELM Classic - HellaSwag: leaderboard
The "HellaSwag" 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. 32 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | text-davinci-003 | 82.2 | #1146 |
| 2 | text-davinci-002 | 81.5 | #1225 |
| 3 | OPT-175B | 79.1 | #1538 |
| 4 | davinci | 77.5 | #1553 |
| 5 | OPT-66B | 74.5 | #1603 |
| 6 | gpt-neox-20B | 71.8 | #1600 |
| 7 | text-curie-001 | 67.6 | #1632 |
| 8 | gpt-j-6B | 66.3 | #1609 |
| 9 | text-babbage-001 | 56.1 | #1648 |
| 10 | text-ada-001 | 42.9 | #1655 |
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-hellaswag · How It Works · Data refreshed daily, snapshot 2026-10-20.