When I switch models in Perplexity and run the same question again, I’m not sure whether model choice and web search are handled separately. Does the selected model affect which pages are found, or mainly how the results are interpreted and written up? If you’ve compared answers this way, did you notice the source set changing too?