Perplexity: Sonar Pro Search

Perplexity Sonar Pro Search is a text-and-image input model built around live web search, offering a 200,000-token context window and up to 8,000 completion tokens per response. It supports reasoning but does not support tool calls or structured output, which narrows its fit for agentic or schema-dependent workflows. Its core proposition is search-augmented generation rather than raw language modeling. On pricing, it runs $3.00 per million input tokens and $15.00 per million output tokens, placing output costs on the higher end of the mid-tier range. There is currently no independent benchmark coverage to evaluate quality against peers, so buyers have no third-party signal to verify how its outputs hold up. It is best shortlisted by teams whose primary need is grounded, citation-backed responses rather than complex integrations, and who are willing to accept an unproven benchmark record in exchange for built-in search capability.