BabyVision: leaderboard
Visual reasoning tasks that 3-year-old children can solve - fine-grained discrimination, visual tracking, spatial perception, and pattern recognition. 388 questions across 22 subtypes. Human baseline: 94.1%.
Metric: Score (%). Source: unipat.ai. Status: years away from saturation. 29 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Median Human | 94.1 |
| 2 | GPT-5.6 Sol | 88.9 |
| 3 | Kimi K3 | 85.7 |
| 4 | Qwen 3.8 27B | 85.6 |
| 5 | GPT-5.5 | 83.6 |
| 6 | Qwen 3.8 Max | 82 |
| 7 | Claude Opus 4.8 | 81.2 |
| 8 | Muse Spark 1.1 | 76.3 |
| 9 | Seed 2.1 Pro | 73.7 |
| 10 | Qwen 3.7 Plus | 70.4 |
| 11 | Kimi K2.6 | 68.5 |
| 12 | Seed 2.1 Turbo | 62.9 |
| 13 | Seed 2.0 Pro (High) | 60.6 |
| 14 | GLM-5.3 Flash | 53.4 |
| 15 | Gemini 3.1 Pro (Preview) | 51.6 |
Interactive version: theaggregate.ai/benchmark?slug=babyvision · How It Works · Data refreshed daily, snapshot 2026-09-05.