Generative Search Benchmarks & Retrieval Datasets
We conduct monthly prompt evaluation tests to measure how search models rank and cite pages. Explore our latest empirical datasets.
Model Citation Share Dataset
Data compiled from 1,200 search queries across informational, conversational, and commercial intent categories (June 2026).
| Retrieval Model | Avg. Citations / Response | Footnote Click-through | Attribution Match Rate |
|---|---|---|---|
| Perplexity AI (Sonar) | 4.8 | 12.4% | 85% |
| ChatGPT Search (GPT-4o) | 3.2 | 9.8% | 78% |
| Anthropic Claude (Sonnet) | 2.1 | 6.5% | 71% |
| Google Gemini 1.5 Pro | 2.8 | 8.1% | 64% |
Check Your Site's AI Citations Compatibility
Run a real-time crawl audit and review immediate structured schema recommendation logs.
Explore Topic Cluster
Audit entity maps and ranking checkups within our cluster.