DeepSeek R-1, a reasoning model from a Chinese AI startup, is now available on Perplexity to support deep web research. peppermint

DeepSeek R1, the Chinese AI startup’s reasoning model that claims to offer performance equivalent to industry-leading models at a fraction of the cost, is now available on US search engine startup Perplexity AI, and it supports deep web research. Will do.

“DeepSeek R1 is now available on Perplexity to support deep web research. With OpenAI O1 there is a new Pro Search reasoning mode selector with a transparent chain of thought in the model’s reasoning, the company said in a post on X.

“We are increasing the number of daily uses for both free and paid to add more capacity during the day. Stay tuned,” reads the post.

Further explaining that DeepSeek on Perplexity is hosted in US/EU data centers, the company said, “Your data never leaves the Western servers. The open source model is hosted completely independently from China. Your privacy and data security is our priority.

A few hours ago, Perplexity AI CEO Arvind Srinivas congratulated DeepSeek AI for coming to #1 on the App Store and said, “For some time, it was not clear who would beat ChatGPT for the first time. Could manage that was #8, a year ago. Looking forward to using all my models for search, assistants and agents this year.

Meanwhile, reacting to the concern raised by an DeepSeek is *open-source*. None of your data goes to China.”

Srinivas also shared a meme to show how Perpliexity Pro users would react.

DeepSeek-R1, released last week by a small Hangzhou-based startup founded in 2023, is 20 times cheaper to use than the OpenAI O1 model, depending on the task, according to a post on DeepSeek’s official WeChat account, Reuters reported. 50 times cheaper.

Bloomberg reported that DeepSeek R1 is close to or better than rival models in several key benchmarks such as AIME 2024 for mathematical tasks, MMLU for general knowledge, and AlpacaValve 2.0 for question-answering performance.