Did you know that Perplexity AI is basically a massive routing engine? Instead of just sticking to one LLM, it classifies your query by intent. It figures out if you’re looking for a quick fact, heavy logic, or clean code, then assigns the "best brain" for the job. The Stack: • Logic: GPT-5.2 and Claude 4.5. • Speed: Gemini 3 Flash and Mistral. • Search: Proprietary Sonar models (optimised to stop hallucinations). The play here is efficiency. You get top-tier performance without having to manually switch between apps. Are you using it?