Why Perplexity's New Sonar Became The New Speed Demon of AI Search Models

Perplexity has just unleashed its latest creation, Sonar, a turbocharged AI search model built on Meta’s Llama 3.3 70B. This model is not just another pretty face in the AI world; it’s a speed demon capable of processing a jaw-dropping 1,200 tokens per second. Yes, you read that right—1,200 tokens per second. That’s faster than most of us can type “What’s the weather today?” on Google. But Sonar isn’t just about speed; it also matches or even surpasses some of the top-tier models in accuracy, all while being kinder to your wallet. Let’s dive into the details of how Perplexity is making waves in the AI ocean.