Not sure where to start? Take a guided tour.
AI Site Grader
AI Query Scanner
llms.txt GeneratorMore than 700,000 valid comparisons power the study
The model divergence benchmark is built from 797,644 valid comparisons.
This is a large comparison set, not a handful of anecdotal prompts.
The numbers behind it.
- Valid comparisons
- Cross-model recommendation comparisons in the study.
- 797,644
- Reports analyzed
- Distinct reports contributing to the benchmark.
- 44,088
The evidence.
| Measure | Result |
|---|---|
| Valid comparisons Cross-model recommendation comparisons in the study. | 797,644 |
| Reports analyzed Distinct reports contributing to the benchmark. | 44,088 |
Population and method
Built from 797,644 valid comparisons across 44,088 reports and 8 models, covering 6,439,133 model responses in the observed window.
Where the evidence stops
Agreement is measured across recommendation outputs, not across hidden reasoning or retrieval context.
Average agreement can hide large differences between query classes and model pairs.
The study measures overlap, not which answer was objectively “right”.
What to do with this
The disagreement signal is structural and repeatable, not just noise in a small sample.
Cite this fact.
Source study: Model Divergence“The model divergence benchmark is built from 797,644 valid comparisons.”
Same Question, Different AI, Different Answers - Trakkr Research · More than 700,000 valid comparisons power the study
Same Question, Different AI, Different Answers · Trakkr Research · Updated 11 March 2026