Sonar

Sonar is Perplexity's flagship model designed to integrate language understanding with real-time web search and information retrieval. As Perplexity's primary offering, Sonar represents the company's approach to combining traditional language modeling with live internet access, positioning it as a research-oriented AI assistant rather than a general-purpose language model. The model supports both text and image inputs with a 127,072 token context window, allowing it to process substantial amounts of information while maintaining access to current web data. Unlike traditional language models that rely solely on training data, Sonar can retrieve and incorporate fresh information from the internet during conversations, making it particularly valuable for tasks requiring up-to-date knowledge or real-time research capabilities. Sonar is primarily used for research, fact-checking, and information synthesis tasks where access to current information is critical. Its integration of web search distinguishes it from standard language models that have fixed knowledge cutoffs, though this specialization means it may not match the raw reasoning capabilities of models specifically optimized for complex analytical tasks.