Revenue and Scale
Perplexity’s annualized revenue exceeded $750 million in August 2026. This amount rose from under $250 million at the start of 2026. Monthly query volumes reached between 1.2 billion and 1.5 billion in mid-2026. The user base hit 45 million monthly active users during the same period. This growth follows the October 2025 free release of the Comet browser. Perplexity’s annualized revenue jumped from under $250 million at the start of 2026 to more than $750 million by August 2026, a trajectory that supports ongoing discussions regarding a potential $30 billion valuation.
The company added a $200 million funding round in June 2026 at a valuation near $20 billion. Nvidia is in discussions for a new round that could value the startup above $30 billion. Revenue passed $450 million in March 2026, which was a 50% gain in a single month. This follows a period where the company added $150 million in annualized revenue in just 30 days during February 2026. The user base grew as Airtel gave free 12-month Perplexity Pro access to 360 million customers starting in July 2025. The Comet browser has 3 million monthly active users. Perplexity also signed partnership deals with seven publishers including CNN, The Washington Post, and Fortune.
Enterprise and API Costs
Perplexity Enterprise Max costs $325 per seat per month. This price creates a $295 gap against OpenAI’s Team plan, which costs $30 per seat per month. I find the Perplexity API pricing deceptive because per-request fees can double a bill without warning. Most LLM APIs charge only per token, but Perplexity adds fees based on search depth. Developers call the Perplexity API costs ridiculous.
| Tier | Monthly Price |
|---|---|
| Free | $0 |
| Pro | $20 |
| Max | $200 |
| Enterprise Pro | $40 per seat |
| Enterprise Max | $325 per seat |
API costs involve three variables: model selection, token volume, and request type. One query with 600 input tokens and 400 output tokens might incur additional costs from search or reasoning operations. For example, a single Deep Research query can cost between $0.30 and $1.30. This cost includes citation tokens and reasoning tokens. Developers who route 50,000 queries through Sonar Pro at high context spend $1,500 per day. This amount is much higher than the $300 per day spent on Sonar at low context. OpenAI’s Team pricing is $25 per seat when billed annually or $30 when billed monthly. Gemini 2.0 Flash relies on token-based pricing.
The Sonar model family includes different tiers. Sonar handles high-volume retrieval. Sonar Pro offers higher token and compute costs for deeper reasoning. Sonar Reasoning Pro is the most expensive tier because it uses multi-step search. A developer using Sonar Pro might pay $14 to $22 per 1,000 queries for the agentic multi-step reasoning capability. Perplexity Pro users also have access to GPT-5 and Claude 4.5 Sonnet.
Accuracy and Research
Perplexity maintains a ~7% hallucination rate. This compares to the 88% error rate some researchers found in Gemini 3 Pro during certain tests. Perplexity provides source citations for every response. This architecture helps researchers verify facts. The Comet browser has 3 million monthly active users. In a 100-query test across five categories, Perplexity scored 74 out of 100. This was higher than the 72 out of 100 achieved by Google.
Perplexity scored 14 out of 20 for factual lookups. For current events, the score was 16 out of 20. While other categories lagged, research synthesis reached an impressive 19 out of 20. However, Google won the local information category with a score of 18 out of 20. Perplexity scored only 8 out of 20 in that category.
You should track your own AI referral traffic instead of relying on these industry estimates. Will Google’s deep integration eventually overcome Perplexity’s citation advantage? Perplexity is the better tool for research and fact-checking.
