DeepSeek V4 Pro API prices cut 75%

DeepSeek permanently lowers V4 Pro API pricing

DeepSeek says it will cut the price of its V4 Pro API by 75%, setting new costs at $0.435 per 1M input tokens and $0.87 per 1M output tokens. The company’s move is designed to make a discount that was previously scheduled to expire on May 31 permanent.

This matters because API pricing is one of the most visible levers in the competitive AI model marketplace. Lower per-token costs can shift incentives for developers building customer-facing applications—especially those where usage volume is tied to responsiveness, conversational depth, or agentic workflows that generate additional tokens.

When a provider makes a discount permanent, it effectively resets expectations for budgeting and product unit economics. For startups and enterprises, that can reduce the financial friction of experimenting with higher-context prompts, larger numbers of tool calls, or more iterative reasoning steps—changes that can otherwise quickly drive up inference bills.

The announcement also fits a broader pattern in model competition where pricing moves are used to win mindshare and grow usage before other feature differentiators fully land. Even without details about capacity or service-level changes, token pricing alone can accelerate adoption.

From a “why it’s news” perspective, the key point is not just that there’s a temporary promotion—it’s that DeepSeek is explicitly extending the discount beyond the original expiration window. That signals an intent to compete on cost alongside performance, and it raises the pressure on other API providers to respond either with their own pricing adjustments or with alternative value propositions.