Chinese AI startup DeepSeek has rolled out discounted off-peak pricing for developers using its AI models, with savings of up to 75%, according to its website.
The Hangzhou-based company stated that between 4:30 PM GMT and 12:30 AM GMT (11:30 AM to 7:30 PM ET), developers will pay significantly less for access to its API, which enables app and web service developers to integrate DeepSeek’s AI models. The Deepseek-chat model (DeepSeek-V3) will be up to 50% off, while the Deepseek-reasoner model (DeepSeek-R1) will see up to a 75% discount.
DeepSeek has been rapidly expanding its AI offerings, with reports suggesting it is pushing for an early launch of its next large language model. The company gained global attention in January with its R1 model, which demonstrated how LLMs could be built on tight budgets without human supervision. Following its release, U.S. tech stocks tumbled on Jan. 27, with Nvidia (NASDAQ:NVDA) losing approximately $589 billion in market cap.
The AI price war has intensified since DeepSeek’s emergence, with Microsoft (NASDAQ:MSFT)-backed OpenAI planning price cuts, and Alphabet (NASDAQ:GOOG) introducing tiered access for its Gemini AI model.
This article first appeared on GuruFocus.