SVLAT - Animation: leaderboard
Scientific Visualization Literacy Assessment Test (SVLAT) multiple-choice questions about animated scientific visualizations; accuracy is averaged across the authors' ten raw model runs.
Metric: Accuracy (%). Source: github.com. Saturation forecast: Around January 2027. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 82.86 | #58 |
| 2 | GPT-5.4 | 73.57 | #79 |
| 3 | Qwen 3.5 9B | 67.14 | #379 |
| 4 | Claude Opus 4.6 | 59.29 | #63 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=svlat-animation · How It Works · Data refreshed daily, snapshot 2026-10-20.