Saturday, March 1, 2025

China’s DeepSeek R1 reasoning model is now on Perplexity AI: Aravind Srinivas shares

Perplexity AI boss Aravind Srinivas has shared that the AI platform will now give access to ‘the world’s most powerful’ reasoning model, DeepSeek R1, developed in China. Sharing on X, Srinivas posted that the model has been made available or Perplexity AI users and updates will be coming soon.

Srinivas’ choice of the term ‘the world’s most powerful reasoning model’ was questioned by an X user, to which he asserted that he believes it is the most powerful reasoning model with transparent reasoning traces. “If I am wrong, tell me what is”, he asked.

DeepSeek R1 was released on January 20, 2025, taking the world by storm with reports of its low-cost development, and open-source availability. DeepSeek R1 is an advanced AI model that has been found to be very good at understanding and generating human-like text, answering questions, and even coding. DeepSeek is an AI company based in Hangzhou, China. It is solely owned by a hedge fund company named High-Flyer, whose founder Liang Wenfeng is also the CEO of DeepSeek.