Perplexity Sonar Pro
Sonar Pro
A model with the live web wired in, and every claim comes back with a source attached.
Updated
The read
Sonar Pro answers from a live search index rather than from training data alone, and returns citations as part of the response instead of as an afterthought. For anything that has to be current, or defensible, that is a different product from a frontier model with a search tool bolted on. The retrieval step costs time, the model underneath is not frontier-class, and you cannot host any of it yourself.
Where it shines
- Live web retrieval built into the model, not bolted on
- Citations returned as structured data with every answer
- Search costs are included rather than billed separately
Worth knowing first
- Slower than a plain model, because retrieval happens first
- The underlying model is not frontier-class on its own
Index score
71.6
Provisional. This score is read from the product's public capability, not from the one-prompt rebuild the fully reviewed entries went through, so treat it as a placing rather than a verdict.
- Rank
- 47 of 53
- Builds
- LLMs
- Category
- Reasoning
- Output
- Text in, cited answers from the live web
- Pricing
- Paid only
- Best for
- Answers that must be current and must cite a source
Scores are our own editorial judgement, weighted craft 55, speed 10, control 15 and value 20.
Also worth a look
Full index- 02GPT-5.6 SolGPT-5.6 SolProvisionalOpenAI's deepest thinker: the one to hand a problem you cannot solve yourself.Reasoning modelPaid only88.6
- 07DeepSeek V4 ProDeepSeek V4 ProProvisionalA reasoning model at the frontier, published under MIT, priced like an afterthought.Reasoning modelFree tier86.8
- 26Hunyuan HY3Hunyuan HY3ProvisionalSwitches between fast answers and slow thinking, and charges almost nothing for either.Reasoning modelPaid only80.0