On May 31, 2026, the promotional discount on DeepSeek V4 Pro's API expires. But DeepSeek has made clear it won't be raising prices back to where they were. In other words, the 75% discount is hardening into permanent list pricing. For a service that was billing one million won a month in API costs, the same usage will now cost 250,000 won. This isn't a hypothetical. It shows up on the first June invoice.

On Hacker News, the developer community, the announcement drew 244 points and 145 comments. What was striking was that a large share of those comments weren't simply celebrating the discount—they described developers who had already begun redesigning their own service architectures around this price. In real terms, the baseline cost of an AI API had dropped another notch.

The Promotion Ended, and the Price Didn't Go Up

The discount on DeepSeek V4 Pro's API pricing was announced from the start as a temporary measure. The expiration was spelled out clearly: 3:59 p.m. UTC on May 31, 2026. Most users expected prices to revert to normal after that point.

In mid-May, DeepSeek announced a different decision through its official documentation and social channels: even after the promotion ended, it would formally lock in the quarter-scale price as permanent list pricing.

Setting that number alongside other major models reveals where it sits. As of May 2026, OpenAI's GPT-4o costs $5 per million input tokens and $15 per million output tokens. Anthropic's Claude Sonnet line runs around $3 for input and $15 for output. Google's Gemini 1.5 Pro offers a free tier up to a certain volume and charges beyond that. Compared directly with these, DeepSeek V4 Pro's permanent fixed price is markedly lower.

On performance, DeepSeek V4 Pro posts numbers competitive with the top models from OpenAI and Anthropic on public benchmarks for coding, mathematics, and complex reasoning. The price has come down even as the performance benchmarks sit near the top.

Behind DeepSeek's ability to hold this price is a difference in architectural design. When DeepSeek R1 was released in early 2025, claims circulated that its training cost was a tiny fraction—tens of times less—of OpenAI's. The exact figures were never independently verified. Even so, the strategy of building price competitiveness on a low inference-cost structure has been consistently maintained. With Google, Amazon, and Meta all continuously lowering the prices of their own AI services, DeepSeek's move can be read as an effort to lock in the floor price first in a competitive field.

Draw in API users with a low price, and once they begin designing services on the assumption of that platform, you arrive at a structure where raising prices later won't easily drive them away. That's why this price freeze is more than a simple extension of a promotion.

The Concerns Raised by a Price This Low

There are reasons this change is hard to welcome without reservation.

DeepSeek is a company headquartered in China. There isn't enough transparency about which servers the data passing through its API travels across, or how that data is handled. Some public institutions in Europe and some companies in the United States have already restricted DeepSeek internally or issued guidance to avoid its use. How this API can be applied under Korea's Personal Information Protection Act and the data-handling rules governing the financial and medical sectors is a matter that requires separate legal review.

The pattern of capturing a market with low prices and then changing the terms has recurred throughout the cloud industry. AWS courted startups with low introductory pricing and free tiers, then adjusted prices in stages once market dependence had grown. When a service that leans heavily on the cloud faces a price increase, switching costs make it hard to move. There is currently no guarantee that DeepSeek won't follow the same path.

Service availability also needs verifying. During the period in 2025 when DeepSeek's service was growing explosively, there were cases where the API became unstable under surging traffic. Adopting it without directly testing real latency and availability in a production environment is a separate risk.

These concerns don't apply with equal weight to every situation. The size of the risk shifts with the nature of a service's data and its regulatory environment. Skipping the review simply because concerns exist, and adopting without checking simply because the price is low, each generate costs in their own different way.

Now Is the Time to Pull Out Your Invoice

The people for whom this price change actually matters are those already wiring an AI API into real operations. If you're a solo operator in Korea, a solo PM, or a middle manager driving an AI rollout, there are things to check right now.

Start by figuring out which APIs you're using and at what volume. More often than you'd think, this gets lumped together as a “ChatGPT subscription” or “AI tool costs.” Without knowing, item by item, which model you're using for which purpose, it's hard to explore alternatives. Pulling out the last three months of invoices and breaking them down line by line is the starting point. If, for example, you're handling both customer-email summaries and social-media draft copy with the same model, you can begin by calculating what happens to your costs when you split off the repetitive, less complex work—like drafting—onto a cheaper model.

Not every task needs the highest-performing model. Teams have reported cutting costs by 30 to 60 percent by routing repetitive work—simple classification, keyword extraction, draft summaries—to cheaper models, and reserving the expensive models for complex reasoning or the final, customer-facing output. Now that prices have come down is also the moment to redesign this structure.

To judge whether to adopt DeepSeek, you first have to sort out the nature of the data you handle. Internal workflow automation, content generation based on public information, and processing tasks that involve no personal data carry relatively low risk. If you're handling customer personal data directly, are contractually required to guarantee where data is processed, or operate in a heavily regulated sector like finance or healthcare, legal review has to come first.

When smartphones first went mainstream, the teams that explored the market and designed services in the first two or three years after the app store opened took favorable positions in the competition that followed. The same outcome repeated when cloud services became a real option for startups. The people who got ahead in these shifts shared something in common: they had accumulated hands-on experience experimenting with exactly how a new environment connected to their own way of working. Knowing how to use a technology and knowing which technology to connect, when, and how are entirely different capabilities.

The fact that AI API costs have hardened at a quarter of their former level is a signal that the window for making that judgment has opened again. If you're working right now without a grip on the cost structure of your own tools, that ignorance is quietly draining unnecessary spending from your invoice every month.