The AI models we track
SimplyRank tracks four AI models — the four that matter most for brand discovery today. Each has different ranking behaviour, so the same prompt can give very different answers depending on the model.
The four engines
ChatGPT (OpenAI)
The largest user base of any AI model. The default for most consumers.
| Detail | Value |
|---|---|
| Provider | OpenAI |
| Model version (current) | GPT-4o |
| Citation style | Inline references, sometimes with linked sources |
| Strength for tracking | Reflects how mainstream consumer queries are answered |
| Pattern to watch | Tends to favour brands with strong consumer presence and broad mentions |
Claude (Anthropic)
Strong editorial discipline. Tends to cite trusted publishers and reason carefully before recommending.
| Detail | Value |
|---|---|
| Provider | Anthropic |
| Model version (current) | Claude Sonnet 5 |
| Citation style | Explicit citations of named sources, often editorial publications |
| Strength for tracking | Reflects how research-led B2B users get recommendations |
| Pattern to watch | Editorial citations and authoritative third-party content move the needle |
Perplexity
Citation-led by design. Always shows source links. Built for people researching purchases.
| Detail | Value |
|---|---|
| Provider | Perplexity AI |
| Model version (current) | Sonar Pro |
| Citation style | Always cites — every claim has a clickable source |
| Strength for tracking | The most transparent picture of which sources are influencing AI answers |
| Pattern to watch | Freshness matters — Perplexity weights recent content heavily |
Gemini (Google)
Google's AI model. Surfaces in the Gemini app and Google AI Overviews.
| Detail | Value |
|---|---|
| Provider | |
| Model version (current) | Gemini 2.5 Flash |
| Citation style | Mixed — some inline citations, some prose mentions |
| Strength for tracking | The closest tie to traditional Google SEO behaviour |
| Pattern to watch | Brands with strong Google SEO often translate visibility into Gemini |
Note on Gemini availability: Gemini has been live across all plans since April 2026.
When a model changes
AI companies retire and replace their models. When one of the four engines moves to a new model, its answers can change overnight even though nothing about your brand did. We list each change here and mark it on your charts with a dashed amber line and a note, so a jump on that day reads as a model change, not as a change in your visibility.
| Date | Engine | From | To | What you'll see |
|---|---|---|---|---|
| 6 July 2026 | Claude | Claude Sonnet 4 | Claude Sonnet 5 | Anthropic retired Claude Sonnet 4 on 15 June 2026, so there are no Claude answers from 16 June 2026 to 5 July 2026. From 6 July 2026, in the workspaces we checked, Claude named more brands per answer and wrote far more links than before, so Claude visibility and citations can jump on that day. |
A chart shows the line only when the dates you've chosen include the change and that engine is in view. To compare like with like, choose dates on one side of it.
Which models your plan includes
| Plan | Models scanned |
|---|---|
| Trial | All four |
| Starter Lite | All four |
| Starter | All four |
| Pro | All four |
| Advanced | All four |
| Agency plans | All four |
Model coverage is not a plan differentiator: every plan, including the trial, scans ChatGPT, Claude, Perplexity and Gemini. What changes between plans is how many prompts you can run against them and how often.
Why we track all four
Visibility isn't consistent across models. A brand that ranks well in ChatGPT can be invisible in Claude. The reasons differ:
- ChatGPT may have been trained on different content than Claude.
- Perplexity actively searches the web at query time and weights freshness.
- Gemini ties closely to Google's index.
Tracking all four lets you spot variance, and variance is what turns into a lever. For example: "why is Claude underweighting us when ChatGPT mentions us in 80% of prompts?" is a question you can only ask when you can see both engines side by side.
What we don't track (yet)
- Bing Chat / Microsoft Copilot. Lower volume than the four we cover. We may add this in future.
- Open-source models (Llama, Mistral, etc.). These don't have a consumer-facing search surface that influences brand discovery.
- Region-specific models (Yandex, Baidu, etc.). On the roadmap if there's customer demand.
What's next
Still need help? Email us at [email protected]. Or browse all help articles.