FlagshipPerplexity

Sonar Pro Search

Sonar Pro Search is Perplexity's flagship search-augmented AI model that combines real-time web search with multimodal understanding across a 200K token context window.

Context 200K
Tier Flagship
Modalities text, image
Input from
$3.00 / 1M tokens
across 1 provider

API Pricing

ProviderInput / 1MOutput / 1MUpdated
$3.00$15.004/14/2026

Prices updated daily. Last check: 4/14/2026

Model Details

General

Creator
Perplexity
Family
Sonar
Tier
Flagship
Context Window
200K
Modalities
Text, Image

Capabilities

Tool Calling
No
Open Source
No

Strengths & Limitations

  • Real-time web search integration provides access to current information beyond training cutoff
  • 200K token context window enables processing of lengthy documents and extended conversations
  • Multimodal support allows analysis of both text and image inputs
  • Source citation capabilities provide transparency and verification for generated responses
  • Search-augmented generation reduces hallucination on factual queries
  • Optimized specifically for information retrieval and research tasks
  • Can access and synthesize information from multiple web sources simultaneously
  • No tool calling support limits integration with external APIs and functions
  • Proprietary model with no access to underlying weights or architecture
  • Search dependency may impact response speed compared to pure language models
  • Performance relies on quality and availability of web search results
  • Limited to Perplexity's search infrastructure and indexing capabilities

Key Features

200K token context window
Real-time web search integration
Multimodal input support (text and images)
Source citation and reference linking
Search result synthesis and summarization
Current information access beyond training data
Multi-source information aggregation
Streaming search-augmented responses

About Sonar Pro Search

Sonar Pro Search is Perplexity's flagship model in the Sonar family, designed specifically for search-augmented generation tasks. As a proprietary model from Perplexity, it represents the company's most advanced offering for applications requiring real-time information retrieval and synthesis. The model features a 200,000 token context window and supports both text and image inputs, enabling multimodal search and reasoning capabilities. Unlike traditional language models that rely solely on training data, Sonar Pro Search integrates live web search results into its responses, allowing it to access current information and cite sources. This search-augmented approach makes it particularly effective for questions requiring up-to-date facts, current events, or recent developments. Sonar Pro Search is commonly used for research assistance, fact-checking, news analysis, and applications where accuracy and recency of information are critical. Its combination of large context window and search integration allows it to synthesize information from multiple sources while maintaining awareness of lengthy conversation histories or document contexts.

Common Use Cases

Sonar Pro Search is designed for applications requiring current, factual information and research capabilities. Its search-augmented approach makes it ideal for news analysis, market research, academic research assistance, fact-checking workflows, and customer support scenarios where up-to-date information is essential. The large context window supports lengthy research sessions, document analysis combined with web search, and complex multi-step research tasks. Organizations use it for competitive intelligence, trend analysis, and any application where the accuracy and recency of information outweighs the need for pure generative capabilities or tool integration.

Frequently Asked Questions

How much does Sonar Pro Search cost per million tokens?

Sonar Pro Search pricing varies by provider and usage type. Check the pricing table above for current rates across all available providers offering this model.

What is Sonar Pro Search best used for?

Sonar Pro Search excels at research tasks, fact-checking, news analysis, and applications requiring current information. Its real-time web search integration makes it ideal for questions about recent events, market conditions, or any scenario where information accuracy and recency are more important than pure generative capabilities.

How does Sonar Pro Search differ from traditional language models?

Unlike traditional language models that rely solely on training data, Sonar Pro Search integrates live web search results into its responses. This allows it to access current information, cite sources, and reduce hallucination on factual queries, though it lacks tool calling capabilities that some traditional models offer.