DeepSeek halts account top-ups as AI models rival ChatGPT’s popularity

Surging demand for the Chinese start-up’s latest AI models appears to have exceeded its server capacity

Reading Time:2 minutes

Published: 10:00pm, 6 Feb 2025

DeepSeek has temporarily stopped developers from topping up their accounts to access the Chinese start-up’s artificial intelligence (AI) models, a sign of the overwhelming popularity of its products.

The company on Thursday disabled the top-up button on its platform website. A notice attributed the change to “server resource constraints”, but added that “existing balances can still be used”.

DeepSeek Platform allows users to engage with its application programming interfaces, or APIs, which serve as a bridge to allow developers and apps to interact with the start-up’s AI models.

Hangzhou-based DeepSeek rose to worldwide fame last month after it released the V3 large language model (LLM) and R1 reasoning model, which the company said showed capabilities comparable to products from ChatGPT creator OpenAI, despite being developed at significantly lower costs.

LLMs are AI models trained to understand human language and perform tasks, such as generating text or answering questions. Reasoning models are designed to perform logical reasoning and inference tasks.

The DeepSeek app is a chatbot supported by the company’s advanced AI models. Photo: Reuters

After DeepSeek released its namesake chatbot on January 10, the app’s daily active users (DAUs) surged quickly to surpass 22 million by the end of the month, trailing only ChatGPT, which had 53 million DAUs at the time, according to market tracker Aicpb.com.