CogSense-Bench - Visual Routines: leaderboard
Metric: Accuracy (%). Source: arxiv.org. Saturation forecast: Around 2033. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Flash | 45.3 |
| 2 | Gemma 3 27B | 43 |
| 3 | O3 | 43 |
| 4 | Intern-S1 | 42.6 |
| 5 | Llama 4 Scout | 41.9 |
| 6 | GPT-5.2 | 37.6 |
| 7 | Grok 4 Fast | 37.6 |
| 8 | Claude Sonnet 4 | 34.4 |
| 9 | Ministral 3 14B | 31.2 |
Interactive version: theaggregate.ai/benchmark?slug=cogsense-bench-visual-routines · How It Works · Data refreshed daily, snapshot 2026-09-25.