DeepSek Pangkas API Prices Outside Busy Hours Up To 75% For Developers

DeepSek, announced a large discount on the price of API usage outside of peak hours (photo; x @BedrockOne)

JAKARTA AI startup from China, DeepSek, on Wednesday 26 February, announced a large discount on the price of API usage outside of peak hours for developers who want to use their AI models in building their own products. This move, announced on its official website, has the potential to pressure competitors both in China and in the global market to lower their prices.

DeepSek's low-cost AI model has previously sparked a major sell-off in the global stock market last month, as investors worry that the company's presence could threaten the dominance of current AI market leaders.

According to DeepSek, from 16:30 GMT to 00:30 GMT (00:30 a.m. to 08:30 p.m. Beijing time), the cost of using the API for the R1 and V3 models will fall by 75% and 50% each compared to the normal price.

Although this period is referred to as "outside of peak hours" in China, the clock actually covers the afternoon and evening in Europe and the United States. The cheap but powerful AI DeepSeeek model has rocked the market, causing a decline in the share price of tech companies.

This discount is the latest step in a series of DeepSek breakthroughs that have disrupted the AI industry both at home and abroad. The company is now accelerating the launch of the next R1 model released in January 2025, according to sources familiar with the company's plans.

DeepSek first sparked a domestic AI model price war in May 2024, when it launched V2, a precursor to a model that now supports their global AI assistants. Since the release of its AI assistant last month, OpenAI ChatGPT maker has cut prices, while Google and Gemini have also introduced discount rates.