DeepSeek is doubling down on its strategy to become the budget-friendly option in the AI market. The Chinese startup has permanently reduced pricing for its flagship DeepSeek V4 Pro model by 75 percent, making what was previously a limited-time promotion the new standard rate.
The price cuts are substantial. DeepSeek V4 Pro now costs between $0.003625 and $0.87 per million tokens, down from the original range of $0.0145 to $3.48 per million tokens. The company was originally planning to end this promotional pricing on May 31, 2026, but has now made it permanent.
This aggressive pricing strategy comes just one month after DeepSeek launched its V4 models, which include both Pro and Flash versions. The company positioned these models as ushering in an “era of cost-effective 1M context length,” directly targeting price-conscious users who need to process large amounts of text.
The move puts serious pressure on established AI providers like OpenAI, Google, and Anthropic. For enterprise customers and power users who burn through millions of tokens daily, DeepSeek’s pricing could represent massive cost savings compared to alternatives like OpenAI’s GPT-5 or Google’s recently released Gemini 3.5 Flash.
This price war comes amid existing tensions in the AI industry. Anthropic has previously accused DeepSeek of conducting “distillation attacks” – a practice where one AI model improperly learns from another company’s more advanced system. These allegations suggest that competitive dynamics between Western and Chinese AI companies are intensifying beyond just pricing.
The broader context here is significant for the AI industry’s future. DeepSeek’s willingness to operate at potentially razor-thin margins could force established players to reconsider their own pricing structures. This could accelerate AI adoption across industries that have been held back by cost concerns, while also raising questions about the sustainability of such aggressive pricing in the long term.
