Michael JohnsonOn October 5, 2026, OpenAI began charging for GPT-Rosalind, a specialized model for life sciences research. But the pricing date does not mean the API is available to every developer: the company makes its use conditional on organizational eligibility and approval of a specific internal research task.
The OpenAI API pricing table lists rates of $5 per million input tokens, $0.50 per million cached input tokens, and $25 per million output tokens for the model. Cache write pricing does not apply to this model. These are per-token processing rates, not a ready-made estimate of a project's cost: the total depends on input volume, response length, and how often requests are repeated.
Price does not mean permission to use it
In an update on September 11, OpenAI said in advance that the published prices would take effect on October 5 and that access would be provided to eligible organizations through its trusted access program. That means today's date marks the start of the planned pricing, not a confirmed new model launch or a general opening of the API.
The company's help article describes narrower conditions for API use: eligible organizations may use it for approved internal research tools, workflows, and applications. It also says the API is not currently available for customer-facing products or external commercial applications. Eligible teams may also get access through ChatGPT Enterprise and Codex; the model appearing on the pricing list does not in itself confirm that a particular organization is entitled to access.
What teams should check
The practical steps here are the reverse of a typical API trial. First, the organization should confirm its eligibility and ensure the intended use complies with OpenAI's access conditions. Only then does it make sense to estimate the token budget for an internal research workflow. If the planned product is intended for external customers, the current help article explicitly rules out that use case for the API—even if the team is ready to pay the published rate.
OpenAI positions GPT-Rosalind for tasks such as genomics, medicinal chemistry, and scientific data analysis. The company also publishes results from its own model evaluations; these are the developer's claims and tests, not an independent assessment that the model improves scientific outcomes at real-world organizations. The available sources contain no independent confirmation that all eligible customers are actually being billed today.
Before using the model or budgeting for it, teams should check the current access conditions and OpenAI's dynamic pricing page. This article was prepared for editorial review; no independent assessment of the model's effectiveness or confirmation of billing today has been established.