Epoch Automation Reports — Data Explorers
Agents · 2026-10-08
Mean human rubric score on two data-explorer generation tasks, with task-specific section weights and human-intervention deductions.
Top models (higher is better)
| Model | Score |
|---|---|
| Grok 4.6 | 74.3 |
| Claude Fable 5.1 | 65.5 |
| Kimi K3 | 61.9 |
| Gemini 3.8 Flash | 52.3 |
| GPT-6 Astra | 46.3 |