Not sure where to start? Take a guided tour.
AI Site Grader
AI Query Scanner
llms.txt GeneratorCross-model consensus benchmark
The 8-model benchmark shows partial overlap, low perfect consensus, and a persistent divergence tail.
Top-line agreement metrics across the full 8-model comparison set.
The key numbers.
- Average agreement
- Mean cross-model agreement rate.
- 43.3%
- Perfect agreement
- Only a small share of prompts produce unanimous outcomes.
- 4.0%
- High divergence rate
- Prompts in the 0-25% agreement bucket.
- 14.6%
- Average top 3 overlap
- Average overlap among top-three results across models.
- 2.8
The ranked view.
- 143.3%
Average agreement
The central tendency of cross-model overlap.
- 214.6%
High-divergence prompts
The share of prompts with the weakest overlap.
- 34.0%
Perfect agreement
The rare case where all models converge fully.
- 42.8
Average top-three overlap
Overlap exists, but not enough to treat lists as identical.
What changed
Consensus is partial, not total
Most prompts live in the middle ground rather than perfect agreement or total chaos.
The divergence tail is operationally important
A 14.6% high-divergence rate is large enough to move market share across models.
| Measure | Result |
|---|---|
| Average agreement Mean cross-model agreement rate. | 43.3% |
| Perfect agreement Only a small share of prompts produce unanimous outcomes. | 4.0% |
| High divergence rate Prompts in the 0-25% agreement bucket. | 14.6% |
| Average top 3 overlap Average overlap among top-three results across models. | 2.8 |
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”.
Cite this tracker.
Source study: Model Divergence“The 8-model benchmark shows partial overlap, low perfect consensus, and a persistent divergence tail.”
Mack Grenfell · Cross-model consensus benchmark
Same Question, Different AI, Different Answers · Trakkr Research · Updated 11 March 2026