Jeff Bezos-Backed Perplexity Rolls Out Self-Hosted DeepSeek R1 Model Amid Chinese AI Controversy: CEO Aravind Srinivas Says, 'There's A Lot Of Misconception...'

On Monday, Jeff Bezos-backed Perplexity AI unveiled a self-hosted version of the DeepSeek R1 model.

What Happened: Perplexity CEO Aravind Srinivas shared the development on X, formerly Twitter, saying the model is now available for research purposes.

The world's most powerful reasoning model, DeepSeek R1, with reasoning traces, is now on Perplexity, for supporting your daily deep web research! Enjoy! Limits will be increased during the day. Updates coming shortly. pic.twitter.com/ZDSUKnfmyK

— Aravind Srinivas (@AravSrinivas) January 27, 2025

See Also: DeepSeek’s R1 Is Impressive But OpenAI CEO Sam Altman Says Will Deliver Much Better Models: ‘Looking Forward To Bringing You All AGI And Beyond’

The announcement led to misconceptions, with some fearing that DeepSeek data might be accessible to the Chinese government or that search results could be censored.

Srinivas addressed these concerns, stating, "All DeepSeek usage in Perplexity is through models hosted in data centers in the USA and Europe. DeepSeek is *open-source*. None of your data goes to China."

He also assured users that the DeepSeek R1 model on Perplexity remains uncensored by sharing a screenshot of an unfiltered response as proof.

Uncensored DeepSeek R1 on Perplexity pic.twitter.com/TPmHrf0xWr

— Aravind Srinivas (@AravSrinivas) January 28, 2025

Subscribe to the Benzinga Tech Trends newsletter to get all the latest tech developments delivered to your inbox.

Why It Matters: Earlier, Ross Gerber criticized DeepSeek’s privacy policy, expressing distrust in Chinese data practices.

The White House has also placed DeepSeek under a national security review.

Previously, in an interview, David Sacks, the White House AI and crypto czar said, “There's substantial evidence that what DeepSeek did here is they distilled knowledge out of OpenAI models and I don't think OpenAI is very happy about this.

Srinivas also spoke about it saying, “There's a lot of misconception that China ‘just cloned’ the outputs of OpenAI. This is far from true and reflects incomplete understanding of how these models are trained in the first place.”

There's a lot of misconception that China "just cloned" the outputs of openai. This is far from true and reflects incomplete understanding of how these models are trained in the first place. DeepSeek R1 has figured out RL finetuning. They wrote a whole paper on this topic called…

— Aravind Srinivas (@AravSrinivas) January 28, 2025

DeepSeek has been recognized for its capabilities, reportedly outperforming other major AI models like OpenAI’s ChatGPT. Access to DeepSeek-R1 via API begins at $0.14 per million tokens, approximately 750,000 words. The newest DeepSeek model is said to be most comparable to OpenAI's "o1" model, which costs $7.50 per million tokens.

Photo Courtesy: Shutterstock.com

Check out more of Benzinga's Consumer Tech coverage by following this link.

Read Next:

• OpenAI’s Sam Altman Flip Flops On Trump After $500 Billion Stargate AI Project, Says ‘I Wish I Had Done More Of My Own Thinking’

Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors.