FoundationalASSIST - Knowledge Tracing: leaderboard
FoundationalASSIST (Worden et al., 2026) is 1.7 million ASSISTments mathematics interactions of 5,000 students with question text, the students' actual answers and skill tags. Given a student's history and the next question, the model predicts whether the student will answer it correctly (about 27,700 predictions per model, zero-shot); accuracy in percent, where always predicting correct scores 51.3, higher is better.
Metric: Accuracy (%; whether the student answers the next question correctly, about 27,700 predictions from 500 students' histories; always predicting correct scores 51.3). Source: arxiv.org. Saturation forecast: Estimated already saturated. 4 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-OSS-120B | 56.2 | #699 |
| 2 | Qwen 3 Next 80B A3B (Thinking) | 55.8 | #591 |
| 3 | Qwen 3 Next 80B A3B Instruct | 54 | #763 |
| 4 | Llama 3.3 70B Instruct | 52.3 | #1353 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=foundationalassist-knowledge-tracing · How It Works · Data refreshed daily, snapshot 2026-11-12.