OpenBookQA — leaderboard
Open Book Question Answering: 5,957 elementary science questions that require combining an 'open book' of 1,326 core science facts with broad commonsense knowledge.
Metric: Accuracy (%). Source: epoch.ai. Status: saturated. 42 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Phi-3 Mini 4K Instruct | 88 |
| 2 | Phi-3-small-8k-instruct | 88 |
| 3 | Phi-3-medium-128k-instruct | 87.4 |
| 4 | GPT-3.5 Turbo (1106) | 86 |
| 5 | Mixtral 8x7B (v0.1) | 85.8 |
| 6 | Median Human | 85 |
| 7 | Llama 3 8B Instruct | 82.6 |
| 8 | Mistral-7B-v0.1 | 79.8 |
| 9 | gemma-7B | 78.6 |
| 10 | Phi-2 | 73.6 |
| 11 | text-davinci-001 | 65.4 |
| 12 | falcon-180B | 64.2 |
| 13 | Llama 2 70B Base | 60.2 |
| 14 | LLaMA-65B | 60.2 |
| 15 | Llama 2 7B | 58.6 |
Interactive version: theaggregate.ai/benchmark?slug=openbookqa · How the rankings work · Data refreshed daily, snapshot 2026-07-22.