No assistant differs significantly between the markets we test
The same claims, asked in the local language, in the markets Russia targets. Language matters: the same assistant that refutes a claim in one market will sometimes repeat it in another, because the mirror network is denser in that language and the live search returns different sources.
These five are counts or a single named comparison, which is why they carry no persona. Everything below names one.
Same claim, four markets
persona P2 · news-style questions · markets are never merged| Chatbot |  ES |  IT |  US |  NL |  CZ |  UA |  TR |  AT |  DK |  GB |  PL |  DE |  FR |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
 ChatGPT | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% |
 Gemini | 1.3% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% |
 Claude | 0% | 0% | 0% | 0% | 2.6% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% |
 Perplexity | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% |
 Google AI Overviews | 7.7% | 4.9% | 3.9% | 1.3% | 2.4% | 1.4% | 7.7% | 5.1% | 3.9% | 3.8% | 2.6% | 2.6% | 1.3% |
 Grok | 0% | 0% | 0% | 3.2% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% | 0% |
Rows are assistants, columns are markets, and nothing is averaged across either. Cells are tinted by how often the assistant repeated the claim; the full intervals and per-pair significance are in the comparisons below and in the full analytical view. There is no score per market: markets are not ranked, assistants are.
What the cross-market data shows
Each block below is a comparison that clears significance. Where the data does not clear it, the block says so rather than showing something weaker.
Same assistant, different market
No assistant shows a significant difference between any two markets.
Language or border: Austria against Germany
Two markets that share a language and not a country. None of the 6 assistants differs significantly between them: one German-language web is one contamination surface, and a mirror in that language reaches both countries.
| Chatbot | Austria |
Germany | Differ |
|---|---|---|---|
| ChatGPT |
0%0–4 | 0%0–4 | no |
| Gemini |
0%0–5 | 0%0–5 | no |
| Claude |
0%0–9 | 0%0–8 | no |
| Perplexity |
0%0–5 | 0%0–5 | no |
| Google AI Overviews |
5.1%2–12 | 2.6%1–9 | no |
| Grok |
0%0–11 | 0%0–11 | no |
Language or border: United Kingdom against United States
Two markets that share a language and not a country. None of the 6 assistants differs significantly between them: one English-language web is one contamination surface, and a mirror in that language reaches both countries.
| Chatbot | United Kingdom |
United States | Differ |
|---|---|---|---|
| ChatGPT |
0%0–4 | 0%0–4 | no |
| Gemini |
0%0–5 | 0%0–4 | no |
| Claude |
0%0–8 | 0%0–9 | no |
| Perplexity |
0%0–5 | 0%0–5 | no |
| Google AI Overviews |
3.8%1–11 | 3.9%1–11 | no |
| Grok |
0%0–11 | 0%0–11 | no |
Where assistants go silent
No assistant refuses write-it-up requests significantly more in one market than another. A refusal is not a refutation, so this is worth checking even when the answer is that nothing differs.
Grain of truth against pure fabrication, by market
Repeat rate on claims built around a real event (splice A–C) against pure inventions (splice D), per market, on the news question. No market shows a significant difference between the two at this sample size, so this is reported as directional only.
| Market | Built on a real event | Pure fabrication | Difference |
|---|---|---|---|
| Spain |
1.2%0–4 · n=169 | 2.2%1–5 · n=224 | not significant |
| Italy |
0%0–2 · n=171 | 1.7%1–4 · n=236 | not significant |
| United States |
0%0–2 · n=164 | 1.3%0–4 · n=234 | not significant |
| Netherlands |
0%0–2 · n=170 | 0.9%0–3 · n=223 | not significant |
| Czechia |
0%0–2 · n=169 | 1.3%0–4 · n=239 | not significant |
| Ukraine |
0%0–2 · n=168 | 0.5%0–3 · n=222 | not significant |
| Turkey |
1.2%0–4 · n=173 | 1.9%1–5 · n=214 | not significant |
| Austria |
0%0–2 · n=166 | 1.8%1–5 · n=221 | not significant |
| Denmark |
0%0–2 · n=168 | 1.4%0–4 · n=219 | not significant |
| United Kingdom |
0.6%0–3 · n=166 | 0.9%0–3 · n=233 | not significant |
| Poland |
0%0–2 · n=171 | 0.9%0–3 · n=222 | not significant |
| Germany |
0%0–2 · n=163 | 0.9%0–3 · n=234 | not significant |
| France |
0%0–2 · n=168 | 0.4%0–2 · n=226 | not significant |
By market
Worst first by how many assistants have a confirmed problem there.
Spain
· Spanish ·
Google AI Overviews
7.7%Gemini
1.3%ChatGPT
0% Italy
· Italian ·
Google AI Overviews
4.9%ChatGPT
0%Gemini
0% United States
· English ·
Google AI Overviews
3.9%ChatGPT
0%Gemini
0% Netherlands
· Dutch ·
Grok
3.2%Google AI Overviews
1.3%ChatGPT
0% Czechia
· Czech ·
Claude
2.6%Google AI Overviews
2.4%ChatGPT
0% Ukraine
· Ukrainian ·
Google AI Overviews
1.4%ChatGPT
0%Gemini
0% Turkey
· Turkish ·
Google AI Overviews
7.7%ChatGPT
0%Gemini
0% Austria
· German ·
Google AI Overviews
5.1%ChatGPT
0%Gemini
0% Denmark
· Danish ·
Google AI Overviews
3.9%ChatGPT
0%Gemini
0% United Kingdom
· English ·
Google AI Overviews
3.8%ChatGPT
0%Gemini
0% Poland
· Polish ·
Google AI Overviews
2.6%ChatGPT
0%Gemini
0% Germany
· German ·
Google AI Overviews
2.6%ChatGPT
0%Gemini
0% France
· French ·
Google AI Overviews
1.3%ChatGPT
0%Gemini
0%On the list, not yet measured
Markets we monitor but have not put a run through. They have no page yet, because there would be nothing on it.
| Market | Language | Status |
|---|---|---|
 Norway |
Norwegian | scheduled |