frontierreasoning

Perplexity AI

Search-augmented AI models with real-time web access baked into every inference call.

3
Models tracked
2022
Founded
Sonar Pro
Flagship model
San Francisco
HQ
Visit perplexity.aiDocumentation

About

About Perplexity AI

Perplexity AI has reimagined language model inference by building real-time internet retrieval directly into the model pipeline. Sonar Pro queries the web at inference time and cites its sources — making it ideal for answering questions about current events, recent research, and live data that closed-weight models can't access. The Sonar Reasoning Pro model combines DeepSeek R1 chain-of-thought with live web retrieval.

perplexity.ai

Key highlights

  • Every Sonar model queries the live web at inference time — not a RAG add-on

  • Answers include citations for every factual claim — verifiable by default

  • Sonar Reasoning Pro combines R1 chain-of-thought with live internet access

  • Significantly lower cost than building a custom RAG pipeline

  • Available via REST API for integration into any application

Models

3 models tracked

Context windows, pricing, capabilities, and API identifiers — everything you need to choose the right model for your workload.

Flagship

Sonar Pro

2024

127Kcontext
TextCode

The most capable Sonar model. Searches the live web at inference time and returns answers with citations. Best for complex questions requiring up-to-date information.

  • Real-time web search
  • Cited sources
  • High accuracy on current events
  • Complex queries
In: $3 / M tokens
Out: $15 / M tokens
sonar-pro

Sonar

2024

127Kcontext
Text

A cost-efficient online model for straightforward search queries. Searches the web at $1/M — ideal for high-volume factual question answering.

  • Fast online search
  • Low cost
  • Simple factual queries
  • High throughput
In: $1 / M tokens
Out: $1 / M tokens
sonar

Sonar Reasoning Pro

2025

127Kcontext
TextReasoning

Combines real-time web search with DeepSeek R1 chain-of-thought reasoning. Ideal for complex multi-step research tasks requiring both current data and deep analysis.

  • Online + chain-of-thought
  • DeepSeek R1 based
  • Complex multi-step reasoning
  • Cited answers
In: $2 / M tokens
Out: $8 / M tokens
sonar-reasoning-pro

Use Cases

What Perplexity AI is best for

Each provider has specialisms shaped by their architecture decisions, training data, and API design. Here is where Perplexity AI performs best.

News and current events monitoring
Research question answering
Competitive intelligence
Live data summarisation
Fact-checked content generation

Get started

Start building with Perplexity AI

Visit the official Perplexity AI platform to access API keys, documentation, and pricing details.

Visit perplexity.aiRead the docsBack to all providers