Perplexity, the generative AI startup, has introduced a new API service named Sonar, enabling enterprises and developers to integrate its advanced AI search capabilities into their applications. The launch of Sonar aims to address limitations in current AI systems, which often provide responses based solely on training data, limiting their ability to deliver accurate, real-time information. In contrast, Perplexity’s API connects to the internet, ensuring answers are based on up-to-date and trusted sources.
The company is offering two versions of the Sonar API to cater to different user needs. The base version, Sonar, is designed to be faster and more affordable, while the more premium Sonar Pro is intended for users requiring more detailed answers and handling of complex queries. With these offerings, Perplexity aims to cater to both businesses looking for efficient, cost-effective solutions and those seeking high-quality, authoritative information for tough questions.
One of the key features of Sonar is its ability for developers to customize the sources from which the AI search engine pulls its information. This flexibility allows enterprises to tailor the search results to specific needs, offering greater control over the AI’s responses.
Perplexity’s move to launch the Sonar API expands the availability of its AI search engine beyond its app and website. Notably, Zoom is already integrating Sonar into its platform, using it to power an AI assistant within its video conferencing tool. This AI assistant provides real-time answers to user queries, drawing from web searches with citations, all without users having to leave the video chat window.
The introduction of Sonar also opens a new revenue stream for Perplexity, which has previously relied on a subscription service for its AI search engine. Given the tech industry’s increasing trend toward lowering API access costs, Perplexity claims its Sonar API is among the most affordable in the market. For the base version of Sonar, pricing is set at $5 per 1,000 searches, plus an additional $1 for every 750,000 input and output tokens, making it an accessible option for developers and businesses.
For those with more advanced needs, the Sonar Pro version offers more detailed responses and handles more complex queries. This version is priced at $5 per 1,000 searches, with additional fees for input and output tokens, including a higher charge for the model’s output at $15 per 750,000 words. Perplexity’s goal with Sonar Pro is to provide a robust tool for handling more intricate tasks, with an added benefit of twice as many citations compared to the base version.
Through Sonar, Perplexity is not only broadening the reach of its technology but is also positioning itself as a key player in the growing AI search space, offering affordable, scalable solutions to enterprises and developers seeking reliable, real-time information.
Read Also: Meta Unveils New Video Editing App ‘Edits’ Amid TikTok Ban Developments