OpenAI and the U.S. General Services Administration have announced a new OneGov agreement that broadens access to ChatGPT across the public sector. The offer removes the platform-access fee and discounts consumption-based usage by 50% for eligible federal, state, local and tribal governments. It is scheduled to take effect October 1, 2026, and run through December 31, 2028.
The price structure moves from license to usage
OpenAI says the standard $15-per-user monthly license fee becomes $0 for participating organizations, with no minimum commitment, while usage is billed at a 50% discount. GSA describes the arrangement in similar terms: agencies pay for what they use, with no platform-access fee, minimum orders or spend commitment. That makes the entry price lower, but it also means total spending will depend on how heavily each agency uses the models and how it manages consumption.
The 27-month term gives agencies a stable procurement window instead of another short pilot. GSA says the agreement covers eligible ChatGPT models, including offerings in FedRAMP-authorized environments, and can be ordered through the Multiple Award Schedule as well as supported resellers, cloud marketplaces and other ordering paths.
Eligibility is broader than the earlier federal offer
The change is not just a new price for federal agencies. OpenAI says it is extending the same offer to state, local and tribal governments. GSA’s release also describes access for the federal executive, legislative and judicial branches, giving the arrangement a government-wide scope that reaches beyond the executive-branch focus of earlier OneGov offers.
OpenAI says more than one million government employees already have access through existing agreements and that the expanded eligibility reaches a U.S. public-sector workforce of roughly 23 million people. Those figures describe potential or existing access, not proof that every eligible worker will use the service or that every deployment will deliver measurable savings.
Cyber defense is part of the package
The agreement also pairs broader general access with a cyber-defense offer. OpenAI says every verified government entity will be approved for Daybreak Blue, its advanced cyber-defender access tier, at 50% off standard commercial pricing, with scaled training and enablement. Defenders can request Daybreak Red for advanced vulnerability research, exploit validation and red teaming at standard commercial pricing.
OpenAI frames those capabilities around emergency response, transportation, public health, utilities and other critical systems. The company also points to earlier government examples, including public-health literature reviews, tax-form digitization, unclaimed-property research and fusion-model work. These are OpenAI-reported case studies rather than independent evaluations, so they are best read as examples of intended use and reported experience.
What the deal does not settle
Removing a platform fee does not remove the operational questions around AI procurement. Agencies still need to define which data may enter a service, how outputs are reviewed, which model capabilities are appropriate for a mission, and how token consumption is monitored. GSA says its agreement incorporates AI terms and conditions intended to safeguard government data, while OpenAI says business inputs and outputs are not used to train or improve its models on eligible ChatGPT Enterprise services.
Newsroom analysis
The broader signal is a shift from one-off government pilots toward a standardized, consumption-based procurement layer. That can reduce acquisition friction and make experimentation easier for smaller agencies, but it can also move the main budget risk from buying seats to managing usage. A 27-month term creates time to learn; whether it creates public value will depend on transparent telemetry, human accountability and evidence from deployments beyond vendor-reported anecdotes.
This story is based on announcements from OpenAI and GSA, the two parties describing the agreement. Neither source is an independent audit of model performance, security outcomes or taxpayer savings, and the agreement’s practical impact will become clearer only as agencies begin ordering and reporting results.