Lab
AI Citation Leaderboard
As of June 30, 2026, Citonyx is cited in 0.0% of AI answers across 19 test prompts (56 runs, ChatGPT). That is the honest baseline — the table below shows which domains AI answer engines currently cite for generative-engine-optimization questions, so you can see who is winning the answer box today.
Last updated · Maintained by Citonyx
Overall citation rate
0.0%
0/56 runs
ChatGPT citation rate
0.0%
0/56 runs
Test prompts
19
fixed citation panel
Total runs
56
2–3 runs per prompt
Competitor domains
199
distinct, cited by AI
Last updated
June 30, 2026
grows as runs accrue
Who AI engines cite for GEO topics
Top 25 domains AI answer engines cited across our fixed prompt set, by number of citations.
| # | Domain | Citations |
|---|---|---|
| 1 | blckalpaca.at | 6 |
| 2 | oversearch.ai | 6 |
| 3 | citeflow.io | 4 |
| 4 | linkbuildingjournal.co.uk | 4 |
| 5 | medium.com | 4 |
| 6 | researchgate.net | 4 |
| 7 | semrush.com | 4 |
| 8 | stackmatix.com | 4 |
| 9 | tryprofound.com | 4 |
| 10 | cintra.run | 3 |
| 11 | conbersa.ai | 3 |
| 12 | geocopy.io | 3 |
| 13 | goodie.ai | 3 |
| 14 | gripped.io | 3 |
| 15 | growwisemedia.com | 3 |
| 16 | heliuxdigital.com | 3 |
| 17 | margen.net | 3 |
| 18 | otterly.ai | 3 |
| 19 | peec.ai | 3 |
| 20 | pierview.ai | 3 |
| 21 | procloser.ai | 3 |
| 22 | promptarchitect.app | 3 |
| 23 | scrunchai.com | 3 |
| 24 | searchengineland.com | 3 |
| 25 | seoscore.tools | 3 |
Download the data
The full run records (56 rows) are public. CC BY 4.0 — attribute Citonyx.
Methodology
- Engines. We query AI answer engines (ChatGPT) and record whether Citonyx appears in the answer as a link, quote, or brand mention.
- Repeats. The protocol targets ≥ 3 runs per prompt/engine pair to account for model non-determinism; the current snapshot has 2–3 runs per prompt. Every individual run is stored as its own record.
- ChatGPT channel. Observed via an anonymous, search-forced backend channel (no signed-in account, search explicitly triggered) so results reflect public, logged-out behavior.
- Planned engines. Gemini and Perplexity are on the protocol but not yet measured — no runs are included for them yet, so the figures above describe ChatGPT only.
- No UI scraping. We do not scrape answer-engine web UIs beyond permitted/anonymous access. Answers are not replayable, so we preserve short verbatim excerpts as evidence.
- Fixed prompt set. The citation panel is a fixed set of 19 GEO-related prompts. It is intentionally narrow, so these numbers describe that prompt set — not all of AI search.
- Updates. Data grows as runs accrue. Current snapshot: 56 runs over 1 measurement date (latest June 30, 2026).