OpenAI just introduced a business seat that costs five times its standard one — and the timing is not an accident.

Two things happened at OpenAI within a few weeks of each other, and read together they tell you exactly where the company's head is. First, at the end of July, it cut the price of its newest models, lowering GPT-5.6 Terra by about 20% to $2 per million input tokens and $12 per million output tokens. Then, on August 10, it did the opposite at the subscription layer: it launched a premium ChatGPT Business seat at $125 per user per month, five times the price of the $25 standard seat. Cutting on one side, raising on the other. That is not indecision. That is a company learning to price like a mature software business — and doing it in public, as it builds toward an IPO.

What ChatGPT Business premium pricing actually changes

The new premium tier sits between the standard Business plan and OpenAI's fully custom Enterprise contracts, filling a gap that had been forcing a hard jump. Standard Business seats stay at $25 per user monthly, or $20 on annual billing. The new premium seat runs $125 per user monthly, or $100 annually. For the money, premium buyers get roughly five times the usage of a standard seat and, critically, the removal of the five-hour usage cap that standard users hit when they lean on the product hardest.

That last detail is the tell. OpenAI is not adding a premium tier because power users asked for more features. It is adding one because its most intensive users were slamming into a usage ceiling — and because serving those users costs OpenAI real money in compute. The five-hour cap was a blunt instrument for controlling inference costs on the standard plan. The premium seat replaces that blunt instrument with a price. If you want to use the product heavily, you pay for the compute you consume, packaged as a $125 seat. It is a cleaner alignment of what a customer pays with what they cost to serve.

Why cut model prices and raise seat prices at the same time

The two moves look contradictory only if you think OpenAI has one product. It has two, and they are priced for two different games.

The model API — the tokens developers buy to build their own applications — is a commodity market in a price war. Google is undercutting on cost with its Gemini models, and a wave of capable, cheaper Chinese and open models is compressing margins across the board. In that market, price is the battlefield, and OpenAI is cutting to defend share. Lowering Terra's price is a defensive move in a fight OpenAI cannot afford to lose on cost.

ChatGPT Business is a different animal. It is a packaged product with a workflow, an interface, integrations, and switching costs — the things that let a software company charge for value rather than for raw compute. In that market, OpenAI has pricing power, and the premium seat is it flexing. The company is doing what every durable SaaS business eventually does: segment its users, identify the ones who extract the most value, and build a tier that captures more of it. The token price is where OpenAI competes; the seat price is where it earns.

The IPO subtext

None of this is happening in a vacuum. OpenAI is moving toward a public offering, and the premium tier is explicitly framed as part of a push to grow and diversify revenue ahead of it. That context changes how you should read every pricing decision the company makes this year.

A company preparing to go public needs a revenue story that a skeptical market will believe: not just enormous top-line growth, but growth with a path to healthy margins. Raw API revenue, sold at prices being competed toward zero, does not tell that story well. High-margin subscription revenue from sticky business customers does. The premium seat is a step toward a revenue mix that leans harder on the durable, defensible part of the business and less on the commodity part. It is OpenAI signaling to future public investors that it can monetize like software, not just sell inference like a utility.

Who this helps and who it squeezes

For OpenAI, the premium tier is close to free money in the near term. Its heaviest ChatGPT Business users were already the costliest to serve and the least price-sensitive; giving them a way to pay more for uncapped access converts a cost center into a margin opportunity. It is the rare pricing move that improves both revenue and unit economics at once.

For enterprise buyers, the new tier is a genuine convenience and a subtle escalation. Companies whose teams were hitting the five-hour cap now have a clean way to buy more without negotiating a custom Enterprise contract. But the segmentation also makes the real cost of heavy AI use more visible. A department that thought it was standardized on $25 seats may discover that its actual power users need $125 ones, and the per-head cost of the product quietly quintuples for the people who use it most. Budget planning around AI tooling just got more complicated.

For competitors, OpenAI's split strategy is a signal and a challenge. Google, Anthropic, and Microsoft are all navigating the same tension between commodity token pricing and defensible product pricing. OpenAI showing that it can raise subscription prices while cutting token prices demonstrates a level of pricing confidence in its packaged product that rivals will have to match. The message to the market is that ChatGPT Business is sticky enough to reprice — a claim about brand and switching costs as much as about features.

For the broader AI market, the divergence is the real news. The price of raw intelligence is falling fast — tokens are getting cheaper by the month. The price of intelligence packaged into a product people actually use in their workday is going up. That gap is where the durable businesses of this era will be built, and OpenAI is now operating on both sides of it deliberately.

The takeaway for operators and builders

The lesson is one every software operator should internalize: sell the commodity at commodity prices and the product at product prices, and never confuse the two. OpenAI is not lowering its token prices because it wants to; it is doing so because that market gives it no choice. It is raising its seat prices because that market lets it. The skill is in knowing which of your offerings sits in which category, and pricing each for the game it is actually in.

If you build on top of these models, the practical implication is to watch the two curves separately. Your input costs — the tokens — are on a downward trajectory that is likely to continue as competition intensifies. Your tooling costs — the seats and subscriptions — are on an upward one as vendors segment and monetize their packaged products. Planning your own pricing and margins means tracking both, and recognizing that "AI is getting cheaper" is only half true. The raw ingredient is getting cheaper. The finished product is learning to charge.

The price of intelligence is falling. The price of a product built from it is not. OpenAI just showed, in a single fortnight, that it understands the difference better than almost anyone — and that the road to a defensible public company runs straight through that gap.

Frequently Asked Questions

What is ChatGPT Business premium pricing?

On August 10, 2026, OpenAI launched a premium ChatGPT Business seat at $125 per user per month, or $100 with annual billing. It sits above the $25 standard Business seat and below custom Enterprise contracts, offering roughly five times the usage of a standard seat and removing the five-hour usage cap.

Why did OpenAI raise seat prices but cut model prices?

The two products serve different markets. The model API is a competitive, commodity market where OpenAI is cutting prices to defend share against Google and cheaper rivals. ChatGPT Business is a packaged product with switching costs, where OpenAI has pricing power and can charge for value rather than raw compute.

How does this relate to OpenAI's IPO?

OpenAI has framed the premium tier as part of a push to grow and diversify revenue ahead of a planned public offering. High-margin subscription revenue from sticky business customers strengthens the revenue story OpenAI can present to public-market investors, compared with commodity API revenue sold at falling prices.

What does the premium tier mean for enterprise buyers?

Buyers whose teams hit the standard plan's five-hour cap can now pay for uncapped, higher-usage access without negotiating a custom Enterprise contract. But it also makes the cost of heavy AI use more visible, since a company's most active users may need the $125 seat rather than the $25 one.

Editor's note — sources: OpenAI on the premium ChatGPT Business tier; Yahoo Finance on the IPO context; CNBC on the GPT-5.6 model price cuts; TechTimes on the five-hour usage cap.

Share this article: