More than 700,000 valid comparisons power the study | Trakkr Research
This is a large comparison set, not a handful of anecdotal prompts.
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
The model divergence benchmark is built from 797,644 valid comparisons.
Why it matters
The disagreement signal is structural and repeatable, not just noise in a small sample.
Supporting metrics
| Metric | Value | Context |
|---|---|---|
| Valid comparisons | 797,644 | Cross-model recommendation comparisons in the study. |
| Reports analyzed | 44,088 | Distinct reports contributing to the benchmark. |
Related pages
Continue through the same study cluster.
- how much do models disagree on brand recommendations - Related answer page
- which query types produce the most consensus - Related answer page
- query class agreement tracker - Related tracker page
Data & Sources
- Same Question, Different AI, Different Answers - Flagship study behind this page
- Page JSON - Machine-readable companion file