How DeepSeek Overtook ChatGPT on the App Store: Low‑Cost AI Model Shakes the Industry

In recent days, the domestic large‑model DeepSeek surpassed ChatGPT, topping the free‑app download rankings of Apple’s App Store in both China and the United States, becoming a hot topic in the global tech community.

On January 27, DeepSeek claimed the first position on the China free‑app list. Simultaneously, it leapt from sixth to first place on the US free‑app list, overtaking ChatGPT, Meta’s Threads, Google Gemini, Microsoft Copilot and other generative AI products.

Following its rapid rise, DeepSeek experienced two service interruptions. On January 26, a brief crash occurred, and the company later confirmed a localized service fluctuation that was resolved within minutes. The outage was attributed to a sudden surge in user traffic after the new model’s release, overwhelming the servers’ concurrent capacity. A subsequent status update indicated that the DeepSeek web/API was unavailable and under investigation.

On the same day, Feng Ji, producer of "Black Myth: Wukong," praised DeepSeek as a potential national‑level technological achievement.

Earlier, Luo Fuli, a key developer of the open‑source DeepSeek‑V2 model, was reportedly recruited by Lei Jun with a multi‑million‑yuan salary offer.

The primary reason DeepSeek has attracted attention is its ability to train a large AI model at a remarkably low cost, rivaling OpenAI’s ChatGPT.

DeepSeek’s R1 pre‑training expense was only US$5.576 million, less than one‑tenth of the cost for OpenAI’s GPT‑4o model.

The company announced its API pricing: 1 yuan per million input tokens (cache hit) or 4 yuan (cache miss), and 16 yuan per million output tokens, roughly one‑thirtieth of OpenAI’s o1 operating cost.

Additionally, DeepSeek is open source.

From the emergence of OpenAI to DeepSeek’s surprise breakthrough, the AI era changes daily, proving that anything is possible.