Epoch Automation Reports — Graphic Design
Agents · 2026-10-08
Mean human rubric score on three graphic-design tasks, with task-specific section weights and human-intervention deductions.
Top models (higher is better)
| Model | Score |
|---|---|
| GPT-6 Astra | 73.3 |
| Claude Fable 5.1 | 64.9 |
| Grok 4.6 | 48.9 |
| Kimi K3 | 40.0 |
| Gemini 3.8 Flash | 32.9 |