High-divergence prompts make up 14.6% of the study | Trakkr Research
A meaningful minority of prompts split the models sharply.
Methodology: Built from 797,644 valid comparisons across 44,088 reports and 8 models, covering 6,439,133 model responses in the observed window.
Claim
14.6% of prompts fall into the high-divergence bucket in the model divergence benchmark.
Why it matters
These are often the highest-opportunity prompts for competitive share shifts.
Supporting metrics
| Metric | Value | Context |
|---|---|---|
| High divergence rate | 14.6% | Prompts in the 0-25% agreement bucket. |
Related pages
Continue through the same study cluster.
- which query types produce the most consensus - Related answer page
- are general and best of prompts more volatile than comparisons - Related answer page
- cross model consensus tracker - Related tracker page
Data & Sources
- Same Question, Different AI, Different Answers - Flagship study behind this page
- Page JSON - Machine-readable companion file