OpenAI Offers Some Enterprise Customers Outcome-Based Pricing
Outcome-based pricing structurally shifts who bears the cost of model failures, from enterprise customers to OpenAI. Token billing's unpredictability, where costs scale with attempts rather than results, has made AI spend difficult for finance teams to justify.
OpenAI has begun offering select enterprise customers a pricing model that charges only when AI tasks successfully complete, rather than for every token consumed. The shift transfers the cost of failed attempts from buyer to vendor: under token billing, customers pay for every failed attempt, while under outcome billing OpenAI absorbs those costs. Token-based billing has produced runaway cost problems, illustrated by one developer running a hundred agents in parallel who accumulated $1.3 million in OpenAI tokens across thirty days. OpenAI had been considering this approach as a way to charge a large share of the value of solutions delivered, with selling solutions rather than tokens described as potentially far more lucrative, though the logistics of implementation were noted as complex.
OpenAI has moved from exploring to actively offering some enterprise customers outcome-based pricing, where charges apply only when AI tasks successfully complete.
The shift transfers the cost of failed attempts to OpenAI rather than the customer, a contrast to token billing under which every failed attempt is billed. Token billing's risk is illustrated by one developer who accumulated $1.3 million in charges running a hundred agents in parallel over thirty days.
Rohan Paul argues token pricing is no longer a comparable unit, proposes cost-per-successful-task metric
OpenAI published a token-based rate card for Enterprise customers, moving them from fixed credits to per-token billing across GPT-5.6, GPT-5.5, GPT-5.4, and deep-research models; Codex had already shifted to usage-based billing on April 2, 2026, with heavy agentic workflows triggering separate charges beyond standard seats.
ChatGPT for PowerPoint followed token pricing after August 6, 2026. OpenAI is also exploring task-based pricing, where clients would pay only for completed tasks rather than tokens, described as potentially more lucrative but logistically complex.
I send one email each morning with the stories that moved. If you would rather just read here, that works too.
Subscribe free