Perplexity announced on Tuesday the launch of its enterprise-level API service "Sonar," allowing businesses and developers to integrate its generative AI search capabilities into their own applications. The service is available in two versions: the basic Sonar and the premium Sonar Pro, and it supports customers in customizing the information sources for their AI search engines.
Unlike traditional AI models that rely solely on training data, Perplexity emphasizes that Sonar can connect to the internet in real-time and obtain information from trusted sources, thereby providing more factual and authoritative answers. Notably, the video conferencing platform Zoom has already integrated the Sonar service, enabling its AI assistant to provide answers based on real-time web searches without users having to leave the video interface.
In terms of pricing, Perplexity adopts a highly competitive strategy. The basic Sonar operates on a fixed pricing model, charging $5 for every 1,000 searches, with an additional fee of $1 for every 750,000 words (approximately 1 million tokens) for input and output. The Sonar Pro version offers more detailed answers and additional citations, priced at $5 for every 1,000 searches, $3 for every 750,000 words of input, and $15 for every 750,000 words of output.
It is worth noting that Perplexity claims that Sonar Pro outperforms industry-leading models such as Google, OpenAI, and Anthropic in the SimpleQA benchmark tests. This startup, valued at $520 million, currently has an annual recurring revenue between $5 million and $10 million and recently completed a $73.6 million funding round earlier this month. The launch of the Sonar API is expected to open new revenue streams for the company and further drive business growth.
