Sonar / Sonar Pro is a Search-optimized LLM from Perplexity (United States).
One-line fit: best used for Grounded Q&A with sources.
Key specifications
| Provider | Perplexity |
|---|---|
| Model | Sonar / Sonar Pro |
| Type | Search-optimized LLM |
| Context window | Large |
| Pricing (indicative) | API + Pro plans |
| Company category | AI Search |
| Region | United States |
| Official site | https://www.perplexity.ai |
Benchmarks and public standing
Search answer quality product metrics
Treat scores as directional. SWE-bench, Terminal-Bench, LMSYS Arena, Artificial Analysis, and vendor cards use different harnesses and are not always comparable 1:1. Re-check the latest official model card before decisions.
What this model is good at
Grounded Q&A with sources
This aligns with Perplexity’s broader strengths:
- Cited web answers
- Research browsing UX
- Agentic search features
Limitations and watch-outs
Not primary SWE agent
- Rate limits, regional availability, and data retention policies vary by plan.
- Agent harness quality (Cursor, Claude Code, Codex, custom tools) can change outcomes more than raw model Elo.
- Open-weight availability (if any) is separate from hosted API quality and safety filters.
Ideal users
- Product engineers shipping features that match: Grounded Q&A with sources
- Teams standardizing on the Perplexity ecosystem
- Agent builders who need a Search-optimized LLM
When to pick something else
- Cheaper volume: compare lower tiers from the same lab or open Chinese/EU alternatives.
- Maximum hard SWE thoughtfulness: compare Claude Fable-class and other coding flagships.
- Giant multimodal corpora / Workspace: compare Gemini-class models.
- Self-host / open weights: Llama, Qwen, DeepSeek, GLM, Mistral open lines.
Parent company
Perplexity – Answer engine with citations. Mixes in-house Sonar models with routed frontier models.
Related models from Perplexity
- Model routing (GPT/Claude/Gemini) – Router product
Data is curated for AIForumSphere model directory mid-2026 directional figures.
