Intel / article

From Tokens to Credits to Outcomes: Pricing AI Across Models, Applications, and Cloud Platforms

Source: arxiv.org
Author: Pratik Gupta
Date: Published Oct 23, 2026

Why it matters

has real marginal cost, so seat pricing is giving way to credits, metering, and outcome pricing. The taxonomy and dataset of dated pricing pages help AI product builders choose a billing model.

Key takeaways · AI-distilled

The study codes public pricing pages and documentation across three cohorts: the full Forbes AI 50 (2026), cloud platforms selling AI services chosen by a documented rule, and a frozen set of 9 distinctive designs, with dated primary sources for each offering.
Subscription with a usage quota is the dominant observable model among AI-native companies: 55% of observable rows, or 45-61% depending on how hybrid edge cases are classified. 23 of the 52 AI-native rows publish no pricing.
The author argues hybrid products are better understood as compositions of a packaging anchor, an allowance mechanism and a billable unit than as mutually exclusive pricing categories.
From the data, the paper derives requirements for automated billing platforms: ledger state for credits, attribution pipelines for outcome pricing, and configuration support for frequent pricing changes.

Terms in this piece · Glossary

inference — Running a trained model to get answers — the phase where AI is actually used, as opposed to trained.

Read the source arxiv.org