HELM Lite - OpenbookQA: leaderboard
Metric: Exact Match (%). Source: crfm.stanford.edu. 91 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude 3.5 Sonnet (20240620) | 97.2 |
| 2 | GPT-4 Turbo | 97 |
| 3 | GPT-4o (2024-08-06) | 96.8 |
| 4 | Claude 3.5 Sonnet (20241022) | 96.6 |
| 5 | GPT-4o (2024-05-13) | 96.6 |
| 6 | Qwen 2.5 72B Instruct | 96.2 |
| 7 | GPT-4 (0613) | 96 |
| 8 | Nova Pro | 96 |
| 9 | Claude 3 Opus (20240229) | 95.6 |
| 10 | DeepSeek V3 | 95.4 |
| 11 | Qwen 2 72B Instruct | 95.4 |
| 12 | Gemini 1.5 Pro (002) | 95.2 |
| 13 | GPT-4 Turbo (Preview) | 95 |
| 14 | Jamba 1.5 Large | 94.8 |
| 15 | Gemini 2.0 Flash (Preview) | 94.6 |
Interactive version: theaggregate.ai/benchmark?slug=helm-lite-openbookqa · How It Works · Data refreshed daily, snapshot 2026-09-19.