Sonar Turbo
Sonar Turbo is Perplexity's latency-optimized search model, engineered for applications where speed is the primary constraint and real-time web grounding is still required. It sacrifices some synthesis depth compared to Sonar Large or Huge in exchange for significantly faster response times, making it the go-to choice for user-facing products with tight latency budgets.
It suits conversational interfaces, live search widgets, and any integration where users expect near-instant answers with source attribution. For simple to moderately complex factual queries, Sonar Turbo delivers grounded, cited responses at a pace competitive with ungrounded models.
Key Features
Optimized for minimum-latency real-time web search responses
Source citations maintained despite speed optimization
Handles straightforward to moderately complex factual queries
Designed for user-facing, latency-sensitive product integrations
Cost-effective for high-throughput real-time search workloads
Perplexity's retrieval stack with a faster inference backend
Ideal Use Cases
Live search widgets in consumer applications
Real-time customer support bots needing current data
Autocomplete or instant-answer features in search interfaces
Voice assistants requiring fast factual lookups
High-frequency news or sports score queries
Example Prompts for Sonar Turbo
Technical Specifications
| Provider | Perplexity |
| Category | Search |
| Modality | Text -> Text (web-grounded) |
Frequently Asked Questions
Try Sonar Turbo now
Start using Sonar Turbo instantly — 100 free credits, no credit card required. Access 750+ AI models through one platform.