A Chinese AI startup introduced discounted off-peak pricing for developers seeking to use its AI models to create products, as revealed on its website on Wednesday. This move may compel competitors in China and globally to reduce their prices.
DeepSeek's cost-effective AI models caused turmoil in global equity markets recently, amid concerns that its emergence could challenge current AI market leaders.
Based in Hangzhou, the company announced that from 1630 GMT to 0030 GMT, the expense of utilizing its API would be up to 75% cheaper. During this period, costs for the API of R1 and V3 models would be reduced by 75% and 50%, respectively, as per a table on DeepSeek's website.
Although the company labels this time span as "off-peak" starting at 0030 and ending at 0830 in Beijing, it corresponds to daytime hours in Europe and the United States. DeepSeek's cost-effective yet potent models have previously created a stir in these regions.
The discounts offered on Wednesday are the latest strategic revisions by DeepSeek that have disrupted the AI industry both in China and internationally. The company is reportedly hastening the development of a successor to the January R1 model, sources close to the company have disclosed.
DeepSeek's initial foray into the market caused a stir last May with its open-source approach and competitive pricing when it launched V2, the precursor to the model fueling its globally recognized AI assistant.
Following the introduction of its AI assistant last month, OpenAI, creators of ChatGPT, reduced their prices, while Google's Gemini has introduced discounted access tiers.