OpenAI Cuts API Tier Spending Thresholds
OpenAI replaced numerical API tiers with Build, Launch, and Grow, halving the spending threshold for maximum throughput to $500.
The company replaced its five numbered tiers with Build, Launch, and Grow, halving the payment required to reach standard maximum throughput.
OpenAI updated its commercial API limits this week, retiring its numerical tiers (Tier 1 through 5) in favor of three named brackets: Build, Launch, and Grow. The change appeared in documentation and through the company’s @OpenAIDevs account on X. For engineering teams shipping software to production, the most practical adjustment sits at the high end: unlocking standard maximum capacity now requires $500 in cumulative credit purchases, down from the prior $1,000 threshold.
How the new API tiers work
The updated framework ties request limits and monthly spending caps strictly to total credit purchases on the account. The Free tier remains untouched, offering up to $100 per month in supported regions.
For paid accounts, the three brackets work as follows:
- Build: Unlocked after $5 in total credit purchases. It provides a monthly usage cap of $500, up from the $100 limit under the old Tier 1.
- Launch: Requires $100 in cumulative purchases. It raises monthly capacity to $5,000. Under the retired system, hitting that same cap required $250 in Tier 4.
- Grow: Requires $500 in cumulative purchases. It provides a $200,000 monthly usage ceiling, matching the top capacity of the former Tier 5 at half the previous spend threshold.
Existing active accounts migrated automatically. Teams on Tiers 1 and 2 moved to Build, accounts on Tiers 3 and 4 shifted to Launch, and Tier 5 accounts moved to Grow. Per-token model pricing and custom enterprise contracts remain unchanged.
Elimination of mandatory waiting periods
Beyond the lower cash barrier, OpenAI removed the mandatory calendar delays that previously governed upgrades between tiers. Under the retired structure, advancing to higher throughput tiers required waiting 7, 14, or 30 days after the initial successful payment, regardless of account balance or project needs.
Support documentation and reporting from Mixed News confirm that tier placement now depends strictly on total credits bought. Once an account reaches the required dollar threshold, rate limits update immediately without calendar holds.
Why the change matters to developers
The older delays caused consistent headaches for early-stage companies and software teams running load tests or preparing product launches. Engineering teams frequently slammed into rate caps (requests per minute and tokens per minute), even when they had the budget to fund immediate usage.
This shift removes practical friction at a time when developer platform competition is fierce. Rivals like Anthropic and Google, alongside specialized providers like Groq and Together, have simplified onboarding to capture enterprise workloads. By dropping artificial waiting periods and cutting entry costs for top throughput, OpenAI closes operational gaps that previously drove frustrated developers to test rival APIs.