DeepSeek Permanently Slashes AI Model Prices by 75% to Challenge Global Rivals

Chinese AI startup DeepSeek has announced an unprecedented price cut in a bid to remain competitive in the global artificial intelligence market and challenge other major technology firms. The company has permanently reduced the price of its latest flagship model, “DeepSeek V4 Pro,” to just one-fourth of its previous cost — a massive 75 percent reduction.

Earlier, the discount had been announced as a temporary promotional offer valid until May 31. However, amid intense competition to capture market share, the company has now decided to make the reduced pricing permanent.

Through this move, DeepSeek aims to establish itself as the most affordable alternative in the rapidly growing AI industry.

Token Costs Drop Sharply

According to information published on DeepSeek’s official platform, after the revised pricing structure, customers will now pay between only $0.003625 and $0.87 per one million tokens. Previously, the same volume of tokens cost users between $0.0145 and $3.48.

The company announced this major price reduction only a month after unveiling its “V4 Pro” and “V4 Flash” models. At the time of launch, DeepSeek claimed it was ushering in an affordable era of “1 million context length” AI technology.

Growing Pressure on Major AI Firms

Industry analysts believe the aggressive price cut will benefit enterprise clients and power users the most, especially developers and corporations that consume billions of tokens daily.

At the same time, the move poses a major challenge to expensive and widely used AI models such as OpenAI’s GPT-5 and Google’s recently launched Gemini 3.5 Flash.

Technology experts say DeepSeek’s aggressive “price undercutting” strategy is likely to increase pressure on US-based AI giants. In particular, the decision could further frustrate rivals such as Anthropic, which had previously accused the Chinese startup of improperly extracting data from its Claude AI model.

DBTech/BMT/OR