Sonar Pro Search

About Sonar Pro Search

Sonar Pro is Perplexity's flagship search-augmented model, combining powerful language capability with real-time web search for accurate, cited responses. Unlike traditional language models that rely solely on training data, Sonar Pro actively searches the internet to provide current, factual information with source citations. The model excels at research queries, fact-checking, current events, and any task requiring up-to-date information. Sonar Pro features sophisticated query understanding that determines when and how to search, synthesizing multiple sources into coherent responses. It demonstrates strong performance on accuracy benchmarks, particularly for questions requiring recent information. The model is ideal for research assistants, fact-checking tools, and applications where accuracy and currency matter more than creative generation. For developers building AI systems that must provide reliable, cited information, Sonar Pro offers unique capability that pure language models can't match. It's particularly valuable for professional research, journalism support, and educational applications.

Model Specifications

💰 Real-World Cost Examples

Estimated monthly costs for common use cases

Perplexity Model Lineup

Compare all models from Perplexity to find the best fit

Similar Models from Other Providers

Cross-brand alternatives with similar capabilities

Grok 3

xAI

Input: $3.00

Output: $15.00

Context: 131k

Grok 3 Beta

xAI

Input: $3.00

Output: $15.00

Context: 131k

Claude Sonnet 4.5

Anthropic

Input: $3.00

Output: $15.00

Context: 1.0M

Grok 4

xAI

Input: $3.00

Output: $15.00

Context: 256k

💡 Cheaper Alternatives

Same Brand (Perplexity)

Cross Brand

🚀 Quick Start

Get started with Sonar Pro Search API

from openai import OpenAI

client = OpenAI(
    base_url="https://api.provider.com/v1",
    api_key="YOUR_API_KEY"
)

response = client.chat.completions.create(
    model="perplexity/sonar-pro-search",
    messages=[
        {"role": "user", "content": "Hello!"}
    ]
)
print(response.choices[0].message.content)

Continue Your Decision