OpenAI simplified its API usage tiers from five to three on 6 October 2026, according to its API changelog. The paid tiers are now called Build, Launch and Grow, and organizations upgrade automatically as their total credit purchases reach each tier’s minimum.
Its rate-limits guide sets the thresholds: Build at $5 in total credit purchases with a $500 monthly usage limit, Launch at $100 with $5,000 a month, and Grow at $500 with $200,000 a month. A free tier, for users in allowed geographies, carries a $100 monthly limit.
Business Pill · WHY AN AI PRODUCT CAN STALL
A one-minute explainer of rate limits: a provider caps how many requests and tokens a customer can send per minute, and an app that hits the cap slows or stops. It teaches the general idea only and says nothing about any company in this story.
The key insight: As we read it, OpenAI has made the jump to high volume easy to reach and automatic. Its guide puts the Grow tier, with a $200,000 monthly usage limit, at $500 in total credit purchases, and organizations move up without asking as their purchases reach each minimum. The ladder now has three paid steps instead of five, according to the changelog.
What the Tiers Allow
The guide lists standard rate limits per tier. For Astra, Sol and Terra, tokens per minute rise from 1,000,000 at Build to 4,000,000 at Launch and 40,000,000 at Grow. For Luna they rise from 2,000,000 to 10,000,000 and 180,000,000.
Requests per minute run from 5,000 at Build to 10,000 at Launch for all four models, and at Grow reach 15,000 for Astra, Sol and Terra and 30,000 for Luna, according to the same table.
The guide says these standard limits differ from the Ultrafast rate limits, and that higher tiers generally provide higher rate limits across models.


How the Limits Work
OpenAI says rate limits are set at the organization and project level, not per user, and vary by model. It also says it sets an approved monthly usage limit for each organization, separate from the spend limits customers can configure themselves.
The guide offers two spend controls: a spend alert, which sends a notification while traffic continues, and a hard spend limit, at which affected API requests return a 429 error.
The Steps Between Tiers
On our arithmetic from the published thresholds, moving from Launch to Grow raises the qualifying credit purchases five times, from $100 to $500, while the monthly usage limit rises forty times, from $5,000 to $200,000. From Build to Launch, both the qualification and the limit rise tenfold or more: $5 to $100, and $500 to $5,000.
The changelog entry that announced the change reads: “Simplified API usage tiers from five to three: Build, Launch, and Grow.” The same 6 October changelog also records the release of the Decisions API in beta with gpt-6-luna.
Luna’s tokens-per-minute limit at Grow, 180 million, is 4.5 times the 40 million listed for Astra, Sol and Terra at the same tier, on our arithmetic.
The Structural Read
On our arithmetic, the last step is the steepest. From Launch to Grow the qualifying purchases rise five times and the monthly limit forty times, according to the thresholds OpenAI publishes.
Luna gets the most room. At Grow, the guide lists 180 million tokens per minute for Luna against 40 million for Astra, Sol and Terra, which on our reading points high-volume traffic towards Luna.
Two caps sit side by side. OpenAI says it sets an approved monthly usage limit for each organization, separate from the spend limits customers set themselves, where a hard limit returns a 429 error.
OpenAI API changelog, 6 October 2026
“Simplified API usage tiers from five to three: Build, Launch, and Grow.”
Three Implications
FEWER STEPS TO HIGH VOLUME The guide puts Grow, with a $200,000 monthly usage limit, at $500 in total credit purchases.
LUNA CARRIES THE MOST THROUGHPUT At Grow, Luna’s standard limit is 180 million tokens per minute, against 40 million for Astra, Sol and Terra.
UPGRADES HAPPEN ON PURCHASE OpenAI says organizations upgrade automatically as total credit purchases reach each tier’s minimum.
The Business Engineer Lens
This story maps onto the Business Engineer framework The CFO’s Guide to the Token Economy.
The framework’s starting point: “The cost of intelligence is collapsing while the cost of deploying intelligence is exploding.”
As we read it, OpenAI’s new ladder acts on deployment volume: the pages we read do not list per-token prices, but they show the usage a customer can run each month rising from $5,000 to $200,000 once $500 of credits has been bought.
What Is Not Established
We read the changelog entry and the usage-tiers section of the rate-limits guide. In the pages we read, OpenAI does not list the limits of the five tiers it replaced, so we cannot say how much any limit changed, and we did not contact OpenAI.
The rate-limit figures are OpenAI’s published standard limits as of 7 October 2026; the guide says limits for each organization can be viewed in its settings.
The Bottom Line
OpenAI has cut its paid API usage tiers to three, Build, Launch and Grow, reached at $5, $100 and $500 in total credit purchases, with monthly usage limits of $500, $5,000 and $200,000. At Grow, Luna’s standard limit is 180 million tokens per minute, according to the guide.
94,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. We read the 6 October 2026 entry in OpenAI’s API changelog and the usage-tiers section of its rate-limits guide, as published on 7 October. All figures are OpenAI’s. We did not see the limits of the previous five tiers and did not contact OpenAI. Nothing here is a forecast, and nothing here is financial or investment advice.
Sources: OpenAI API changelog, 6 Oct 2026 entry · OpenAI API docs: Rate limits, usage tiers









