Request for Information on Base Models Used in Perplexity API

[ymuichiro]

I am considering using the Perplexity API, but my organization prohibits the use of DeepSeek models. I would like to know which base models are being used for each API provided by Perplexity, especially for those that perform reasoning tasks.

Previously, llama was explicitly mentioned as a base model, but this information does not seem to be publicly available now.
Could you please share information about which underlying models are used for each Perplexity API offering?

This would help me determine whether I can utilize the API without violating my organization’s compliance policies.
Thank you for your consideration.

• sonar-deep-research (???)
• sonar-reasoning-pro (DeepSeek R1)
• sonar-reasoning (???)
• sonar-pro (???)
• sonar (???)
• r1-1776 (DeepSeek R1)

※ I understand that Perplexity’s DeepSeek is hosted in the United States.

[ymuichiro]

• sonar (Llama 3.3 70b ? https://x.com/perplexity_ai/status/1889392617479082323?s=61)