Sonar Large
Perplexity's flagship hosted model with always-on web search grounding and real-time citations.
Pricing
Input$1 / 1M
Output$1 / 1M
$5 per 1,000 search requests. Pricing for search separate from token usage.
Cost bandBalanced
Free tierNo free API tier; consumer perplexity.ai has free tier.
Limits
Context window200k
Max output8,192
Knowledge cutoffReal-time web
DeploymentPerplexity API
OpenAI-compatible APIYes
ModalitiesText
Where it fits
Best forResearch assistants, news-aware chatbots, and any workflow where up-to-date web data is required.
Business fitUnique position: search-native LLM API — differentiator when knowledge freshness matters.
Capabilities
Real-time web searchCitationsSearch grounding
Benchmark scores
Composite74 / 100
Coding66 / 100 (HumanEval (vendor))
Reasoning74 / 100 (MMLU-Pro (vendor))
Long context78 / 100
Instruction following78 / 100 (IFEval (vendor))
Editorial aggregate of published benchmark results as of Mar 1, 2025. These are not independent measurements run by this site.
Change history
- Feb 22, 2025DeprecatedSonar Large retired Perplexity retired the legacy Llama-Sonar online models on Feb 22, 2025 in favor of the current Sonar line. Perplexity changelog
Data
Last checkedJan 10, 2026
Verify with provider documentation before committing spend. Report an error and it gets fixed with a logged correction.