AI for Education Pedagogy - Science: leaderboard
The science slice of the pedagogy exam set: four-option multiple-choice questions on teaching and assessing science, drawn from Chilean teacher qualification exams and translated into English. Items are answered closed-book in a single turn and scored by exact match on the answer letter. Higher is better. Source: arxiv.org/abs/2506.18710.
Metric: Accuracy (%). Source: www.fab-ai.org. Saturation forecast: Estimated already saturated. 249 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Grok 4.6 | 96.72 | #26 |
| 2 | Qwen 3.5 Plus | 95.08 | #125 |
| 3 | GPT-6 Astra (High) | 95.08 | #6 (GPT-6 Astra) |
| 4 | Gemini 3.1 Pro (Preview) | 94.54 | #58 |
| 5 | Gemini 3 Pro | 94.54 | #80 |
| 6 | Gemini 3.8 Flash | 94.54 | #21 |
| 7 | Qwen 3.6 Max Preview | 94.54 | #85 |
| 8 | GPT-5.5 | 93.99 | #31 |
| 9 | Qwen 3.5 27B | 93.99 | #187 |
| 10 | GPT-6 Astra | 93.99 | #6 |
| 11 | Claude Fable 5.1 | 93.99 | #9 |
| 12 | Gemini 3 Flash | 93.44 | #99 |
| 13 | Qwen 3.7 Max | 93.44 | #75 |
| 14 | Qwen 3.8 27B | 93.44 | #105 |
| 15 | Gemini 3.6 Flash | 93.44 | #51 |
No result here: #3 Claude Opus 5.5, #7 GPT-6.1 Sol, #8 Gemini 4 Argon.
Interactive version: theaggregate.ai/benchmark?slug=ai-for-education-pedagogy-science · How It Works · Data refreshed daily, snapshot 2026-10-20.