FIELD GUIDE / STABILITY AI

Stability AI Image API Credits: Translate Per-Operation Debits Without Mixing Currencies

Stability AI offers a useful example of a developer API priced in proprietary credits rather than a direct USD cost for every operation. The vendor publishes the conversion between credit purchases and list-dollar value, as well as.

Stability AI offers a useful example of a developer API priced in proprietary credits rather than a direct USD cost for every operation. The vendor publishes the conversion between credit purchases and list-dollar value, as well as operation-specific credit debits.

Verification scope: Official sources were checked for the 25 October 2026 cutoff. List tariffs and rights apply only to the documented provider, region, tier, exact model and sales channel. This is an editorial draft, not a verified account quote; unknown effective dates remain null.

One credit has a published list-value relationship

The Stability AI developer-platform price page states a credit purchase ratio of USD 0.01 per credit and describes API usage as credit-based. It also advertises an introductory grant of 25 free credits. A free grant is not a recurring rebate on every month’s bill, and available credits can depend on promotions and account eligibility. The base conversion is specific to Stability credits; it must not be applied to Runway, Luma or ElevenLabs credit balances. Treat purchased credit cost and spent operation credits as two linked but distinct records.

Image services have different tariffs

In the provider’s image table, Stable Image Ultra is listed at eight credits and Stable Image Core at three. Creative Upscaler, object removal, inpainting, outpainting and other edit tools carry separate values; passing the same photograph through three operations creates three potential charge events, not a single image. A hypothetical workflow with 100 Image Core generations and 25 Stable Image Ultra generations uses 500 credits under the quoted rates: 100×3 + 25×8. At USD 0.01 per credit that implies USD 5 of list-equivalent credits before any account-specific tax or purchase conditions.

Edit stages and retries determine accepted-asset cost

An image asset may be generated, enlarged, relit, background-removed and reviewed. A quote for only the first generation cannot predict total spend across these steps. Track successful requests, failed requests, edits, input image requirements and whether the vendor bills unsuccessful operations. If a generated batch yields only ten approved images from one hundred attempts, divide all actual credit consumption by ten accepted outputs. Any stated approval ratio must come from the user’s real sample or be labeled hypothetical, never from fabricated first-hand testing.

Credit top-ups and accounting controls

Financial reporting should preserve how much fiat currency bought the credits and when, then decrement the proprietary balance as API operations consume it. Expiry, refund rules, special promotions and country-specific taxes require contract evidence; the public list ratio alone does not establish them. Never silently describe 500 Stability credits as 500 US dollars. Procurement can compare effective cost per accepted image with other API services only after converting their own meters from first-party rates and the same output-quality requirements.

What is and is not verified

These values are from an official live pricing page checked for the October 25, 2026 archive. The site does not disclose the first effective date of every image-model debit in the extracted reference, so effective_at is left null. Actual availability, model revisions and output rights should be reconfirmed at purchase. The AI 03 team handles imaging techniques and quality comparisons; the pricing archive stores only specific operation, currency and unit identities.

Decision checklist and limitations

Before purchase or deployment, retain the literal model or plan identifier, payer/merchant, country or cloud region, billing currency, exact usage meter, source URL, checked date, effective date if explicitly supported, and any plan commitment or promotional eligibility. Recalculate from the companion price and quota snapshots rather than assuming that a remembered list tariff will remain unchanged. A credit allowance, maximum input length, published token rate and confirmed billed currency answer different questions; never add them as if they were the same quantity.

When comparing another provider or route, use the same input/output distribution and acceptable result definition. Include unsuccessful attempts, optional tools, storage, tax, payment and human review only when those components are actually measured or supported by an explicit source. Vendor marketing and documentation are not independent product tests. This article makes no claim to have personally tested paid endpoints, purchased plans, or checked account-specific taxes or availability.

Sources & Further Reading

CONTINUE EXPLORINGExplore related files ↗