DeepSeek’s iOS App Overtakes ChatGPT on App Retailer, Tech Leaders React to the Open-Supply AI Mannequin’s Rise

DeepSeek for iOS has overtaken OpenAI’s ChatGPT to high the App Retailer’s “High free apps” chart within the US. The eponymous Chinese language firm launched the open-source DeepSeek-R1 synthetic intelligence (AI) mannequin final week, which outperforms OpenAI’s o1 AI mannequin in a number of benchmarks. This launch has made the corporate and its AI chatbot the speak of the city, with a number of Silicon Valley tech leaders reacting to its sudden rise. Notably, the DeepSeek app is solely free to make use of, with no subscription tiers introduced up to now.

DeepSeek Overtakes ChatGPT

Ever since OpenAI launched the ChatGPT for iOS app in Might 2023, it has constantly been among the many high free purposes on the App Retailer and has remained the highest-ranked AI app on the platform. Nevertheless, this throne was taken by the DeepSeek app not too long ago because it reached the highest of the chart. The rise is being attributed to the current launch of the DeepSeek-R1 AI mannequin.

Notably, little is understood concerning the Hangzhou-based AI agency which was based in 2023 and has launched a number of open-source massive language fashions (LLMs). Whereas US-based tech corporations have additionally launched open-source fashions, with Meta being a notable point out locally, none comes near the capabilities and scale of what the DeepSeek fashions provide. The corporate additionally claimed to have constructed the mannequin with a value of $6 million (roughly Rs. 51.8 crore), considerably cheaper than AI fashions of this scale.

A number of Silicon Valley-based tech leaders have reacted to the sudden rise of those fashions. Enterprise capitalist Marc Andreessen referred to as the R1 AI mannequin “probably the most superb and spectacular breakthroughs I’ve ever seen,” whereas Perplexity AI Co-Founder and CEO Aravind Srinivas congratulated the Chinese language agency for changing into the primary AI app to beat ChatGPT.

In a LinkedIn publish, Yann LeCun, one of many godfathers of AI and the present Vice President and Chief AI Scientist at Meta, stated “Open supply fashions are surpassing proprietary ones.” Holger Zschäpitz, a senior monetary reporter at German free-to-air tv information channel Welt, argued DeepSeek “might symbolize a giant menace to US fairness markets.”

DeepSeek Is Not the Solely One

Whereas DeepSeek has damaged the pattern of corporations open-sourcing smaller and fewer succesful AI fashions whereas holding the frontier fashions beneath a proprietary paywall, it isn’t the one one. One other Chinese language agency, Kimi AI, introduced the discharge of the Kimi k1.5 AI mannequin. The corporate claims it’s an “o1-level multimodal LLM that outperforms GPT-4o and Claude Sonnet 3.5 on a number of benchmarks”.

Notably, Kimi AI has made the online model of its chatbot fully free to make use of with limitless utilization, much like DeepSeek. The AI mannequin also can carry out real-time internet searches, analyse as much as 50 information throughout a various vary of codecs, and has picture understanding capabilities as properly. Whereas its technical report is out there on GitHub, it’s presently not obtainable in open-source.