GeoDial: leaderboard
Several educational domains rely heavily on diagrams and visual cues, yet most existing tutoring datasets are limited to text-only interactions. This limits the development of AI tutors that can teach in visually grounded ways used by human instructors. Thus, we introduce GeoDial, a multimodal tutoring dataset of over 1.3K teacher-student dialogs in the domain of geometry collected from experience
Metric: Act Macro-F1 (self-reported). Source: benchmarklist.com. Saturation forecast: Around October 2026. 18 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4 | 32.7 | #199 |
| 2 | Gemini 2.5 Flash | 30.4 | #245 |
| 3 | GPT-4o | 30 | #343 |
| 4 | Qwen 3 32B | 29.5 | #443 |
| 5 | Qwen 3 8B | 27.1 | #674 |
| 6 | Llama 3.1 8B | 17.1 | #1151 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=geodial · How It Works · Data refreshed daily, snapshot 2026-10-20.