MathTutorBench: leaderboard
Open-ended math tutoring benchmark for measuring pedagogical capabilities such as mistake diagnosis, Socratic questioning, and scaffolding.
Metric: Overall Average (imported). Source: eth-lre.github.io. Source: benchmarklist.com. Saturation forecast: Estimated already saturated. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o | 63.78 | #343 |
| 2 | Llama 3.1 70B Instruct | 55.22 | #561 |
| 3 | Llama 3.1 8B Instruct | 47 | #1035 |
| 4 | Llama 3.2 3B Instruct | 46.89 | #1352 |
| 5 | Qwen2.5-Math-7B-Instruct | 31.67 | #1444 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=mathtutorbench · How It Works · Data refreshed daily, snapshot 2026-10-20.