Perplexity has launched its Sonar API, introducing real-time, citation-backed AI search functionality tailored for enterprises and developers. By providing affordable and customizable search tools, the company aims to strengthen its position in the AI search market.
The Sonar API, which debuts with two pricing tiers—Sonar and Sonar Pro—marks a significant expansion of Perplexity’s offerings.
The API empowers developers to integrate generative AI-powered search features into their applications, with capabilities ranging from fast and cost-effective solutions to advanced query handling for complex use cases. The announcement follows a $500 million funding round that values the company at $9 billion.
Customizable Features for Diverse Applications
Perplexity’s Sonar API is designed to meet the needs of developers and enterprises seeking flexibility in their AI-powered search tools. The base tier, Sonar, emphasizes speed and affordability.
Priced at $5 per 1,000 searches, with additional costs of $1 per 750,000 input and output tokens, Sonar offers a lightweight, scalable solution for applications requiring quick and reliable question-and-answer functionalities.
For more complex use cases, the Sonar Pro API tier delivers enhanced capabilities, including a larger context window of 200,000 tokens, support for multi-step queries, and twice as many citations per search compared to the base tier.
This advanced option is priced at $3 per 750,000 input tokens and $15 for the same volume of output tokens. These features aim to improve factual accuracy and provide developers with tools to refine the quality of their outputs.

In a blog post, Perplexity highlighted the API’s flexibility, stating: “To optimize for factuality and authority, APIs require a real-time connection to the Internet, with answers informed by trusted sources. We’re also expanding public access to advanced features like JSON mode and search domain filters for select usage tiers.”
Real-World Applications Highlight Sonar’s Potential
Several organizations have already adopted the Sonar API, showcasing its versatility in diverse industries. Zoom has integrated Sonar Pro into its AI-powered chatbot, allowing users to retrieve real-time answers during video calls without leaving the interface.
Copy.ai, another early adopter, uses the API to streamline sales research. The company reports that its integration with Sonar saves teams up to eight hours per week, improving productivity and increasing throughput by 20%. Similarly, Doximity employs the API to provide doctors with citation-backed answers to medical queries, ensuring accuracy and trustworthiness in high-stakes decision-making.
Competing in the AI Search Market
Perplexity’s Sonar API enters a competitive AI search landscape dominated by major players such as OpenAI, Microsoft, and Google. OpenAI’s ChatGPT Search offers real-time web access powered by GPT-4o.
Google’s Gemini AI and Microsoft’s Bing Generative Search emphasize multimodal capabilities and conversational follow-ups, catering to users seeking comprehensive search solutions.
Related: Meta Joins the AI Search Battle Alongside OpenAI, Google, Perplexity
Perplexity distinguishes itself by prioritizing source transparency and real-time data retrieval. Its API allows developers to customize data sources and adjust key parameters, such as Top P and presence penalties, to reduce nonsensical output and ensure relevance.
The company emphasized this approach in its blog post: “Real-time, citation-backed answers are a must for enterprise-grade tools.”
Benchmarks and Performance
According to Perplexity Sonar Pro offers impressive performance, achieving an F-score of 0.858 on the SimpleQA benchmark, a widely recognized standard for evaluating the factual accuracy of AI models.
This score surpasses results from leading competitors, including Google, OpenAI, and Anthropic. The base Sonar model also performed competitively, scoring 0.773.
The SimpleQA benchmark measures a model’s ability to answer fact-based queries accurately, making it particularly relevant for enterprise users requiring high standards of precision and reliability.
Scaling Revenue and Addressing Challenges
The Sonar API represents a strategic effort by Perplexity to diversify its revenue streams. Previously, the company relied primarily on subscription-based revenue from its AI search engine, which offers features like a built-in code interpreter and one-click e-commerce capabilities through its Shop Like a Pro tool.
With an estimated annual recurring revenue of $5–$10 million, Perplexity’s API launch positions the company for broader adoption and growth.
However, the company faces challenges in scaling its offerings. Data privacy, compliance, and scalability remain critical concerns for enterprise clients. Perplexity AI is accused of accessing and summarizing articles from the New York Times and other publishers without proper authorization, leading to allegations of unethical web scraping and misuse of protected material.