Whiteboard AI Tutor - Blind Review: leaderboard
Whiteboard AI Tutor - Blind Review. Six models produce whiteboard lessons from a frozen Chinese prompt across eight science, math and economics scenarios. Blind Fable5.2 image review of median-linter repeats; mixed low/medium cohort with10or11boards per model. Standard pass and strict geometric-rubric pass percentages retain native denominators; human audit18boards. Not tutoring cost or latency. Metric: Reviewed boards passed (%). Higher is better.
Metric: Reviewed boards passed (%). Source: github.com. Saturation forecast: Estimated already saturated. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 4 Argon | 100 | #11 |
| 2 | Claude Haiku 5.5 | 81.82 | #65 |
| 3 | Gemini 3.8 Flash | 72.73 | #36 |
| 4 | Claude Sonnet 5.5 | 63.64 | #19 |
| 5 | DeepSeek V4 Pro | 50 | #188 |
| 6 | DeepSeek V4.1 Flash | 40 | #59 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=whiteboard-ai-tutor-blind-review · How It Works · Data refreshed daily, snapshot 2026-11-13.