A lightweight, cost-effective search model optimized for quick, grounded answers with real-time web search.

  • Model Type: Non-reasoning
  • Use Case: Ideal for quick searches and straightforward Q&A tasks.
  • Context Length: 128k

Key Features:

  • Real-time web search-based answers with citations
  • Optimized for speed and cost

Real-World Examples:

  • Summarizing books, TV shows, and movies
  • Looking up definitions or quick facts
  • Browsing news, sports, health, and finance content

​
Pricing

Legacy Pricing (Active Until 04/18/2025)

Legacy Pricing (Default if no search mode is specified)

Pricing:

Pricing ComponentCost
Input Tokens (Per Million)$1
Output Tokens (Per Million)$1
Price per 1,000 Requests$5

New Pricing (Available Now)

New Pricing with Search Modes
Explicit API call with search_context_size required to use new pricing.

Pricing ComponentHigh  Medium  Low  
Input Tokens (Per Million)$1$1$1
Output Tokens (Per Million)$1$1$1
Price per 1,000 Requests$12$8$5

​
Test the Model

curl --request POST \
  --url https://api.perplexity.ai/chat/completions \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '{
  "model": "sonar",
  "messages": [
    {"role": "user", "content": "What is the latest news in AI research?"}
  ],
  "max_tokens": 200
}'

Sample Response Metadata

{
  'id': 'ad748e84-a4bb-4665-afc9-40aac3b39e3c', 
  'model': 'sonar', 
  'created': 1743529296, 
  'usage': {'prompt_tokens': 9, 'completion_tokens': 384, 'total_tokens': 393}, 'citations': ['https://news.mit.edu/topic/artificial-intelligence2', 'https://www.sciencedaily.com/news/computers_math/artificial_intelligence/', 'https://aiweekly.co', 'https://techcrunch.com/category/artificial-intelligence/', 'https://www.artificialintelligence-news.com'], 'object': 'chat.completion', 'choices': [{'index': 0, 'finish_reason': 'stop', 'message': {'role': 'assistant', 
  'content': "Here are some of the latest developments in AI research:\n\n## Recent AI Research Developments\n\n### 1. **Improved AI Image Generation**\n- **MIT Research**: An AI tool has been developed that combines two popular methods to generate high-quality images faster and more efficiently than current state-of-the-art approaches. This tool can run locally on devices like laptops and smartphones, using less energy[1].\n\n### 2. **Enhanced Autonomous Learning**\n- **Torque Clustering Algorithm**: Researchers have introduced a new AI algorithm called Torque Clustering, which significantly enhances autonomous learning and pattern discovery in AI systems, requiring minimal human intervention[2].\n\n### 3. **Cyborg Insects Navigation**\n- **Autonomous Navigation**: A research team has developed two new autonomous navigation systems for cyborg insects, enabling them to better navigate through unknown environments using simple circuits[2].\n\n### 4. **AI in Smart Home Security**\n- **AIoT Framework**: Scientists have developed a new AIoT framework, MSF-Net, to enhance smart home security by integrating AI with IoT technology and WiFi, providing robust security solutions[2].\n\n### 5. **Advancements in Large Language Models**\n- **Reasoning Capabilities**: Studies have shown that large language models reason about diverse data types in a general way, similar to human brains, indicating advanced capabilities in handling complex information[2].\n\n### 6. **Regulatory Frameworks for AI**\n- **European AI Act**: The European Union is working on regulatory models like the AI Act to ensure ethical and safe AI use, particularly in data centers, addressing environmental impacts[3].\n\n### 7. **OpenAI Leadership Changes**\n- **Executive Shifts**: OpenAI has reshuffled its leadership, with CEO Sam Altman focusing more on technical aspects while expanding COO Brad Lightcap's responsibilities, signaling strategic adjustments in the company's direction[3]."}, 
  'delta': {'role': 'assistant', 'content': ''}}]
}