OpenAI replaces its $1 government pilot with a 50 percent usage discount for federal agencies.

The discounted arrangement includes no minimum spending requirement, allowing agencies to use the models without committing to a set purchase volume.
Access under the agreement is limited to uses and environments already authorized through the Federal Risk and Authorization Management Program (FedRAMP), with the GSA’s AI terms and conditions providing additional oversight.
When the pilot was initially announced, OpenAI said more than 90,000 users across over 3,500 federal, state and local agencies had sent more than 18 million ChatGPT messages for work-related purposes since 2024.
The original $1-per-agency listing remained visible on the GSA’s OneGov marketplace, with an expiration date of September 30, 2026, even after the replacement pricing was announced; OpenAI’s government webpage had also not yet reflected the change.
Anthropic’s government push became entangled in a dispute over safeguards for surveillance and autonomous strikes in national-security contracts. After the company was designated a supply-chain risk and challenged the resulting ban in court, a preliminary injunction led the GSA to reverse an attempt to remove Anthropic from its central AI platform.
OpenAI is ending its ultra-cheap $1-per-year pilot for U.S. government agencies and switching to usage-based pricing with a 50% discount from regular rates. The new deal takes effect October 1 and runs for more than two years, opening access to federal agencies and eventually state, local, and tribal governments Voice of Emirates. During the pilot, about 3.5 million federal employees used ChatGPT and saved an estimated $1.4 billion.
The shift marks OpenAI's strategy to lock in government customers while undercutting rivals like Anthropic. Unlike the fixed $1 price, the new agreement requires no minimum spending, letting agencies pay only for what they use GuruFocus. The General Services Administration's OneGov platform will handle procurement and ensure all access meets federal security rules.
The original $1-per-agency offer was a loss leader designed to build adoption. Now OpenAI is betting discounts and convenience will keep agencies hooked. The 27-month contract, running through 2027, gives federal workers access to ChatGPT at half the normal commercial rate Voice of Emirates. No minimum purchase means small agencies can experiment cheaply.
When the pilot launched, OpenAI claimed more than 90,000 users across 3,500 federal, state, and local agencies had sent over 18 million ChatGPT messages for work. That usage data gave OpenAI proof that government buyers would stick around — as long as the price stayed reasonable GuruFocus.
Access to OpenAI's models is strictly limited to uses already approved under the Federal Risk and Authorization Management Program, or FedRAMP. This means agencies can't use the discount on new applications without getting formal security clearance first GuruFocus. The GSA adds its own terms and conditions on top to ensure federal oversight.
The limitation is intentional. Federal security officials want to prevent rogue departments from deploying AI in sensitive roles without review. Every use case gets vetted. Every access point gets logged. The result: agencies get cheap AI, but only in ways Washington approves.
Anthropic, OpenAI's rival, has also pushed for federal contracts. But Anthropic got tangled in a clash over AI safeguards for surveillance and autonomous weapons in national-security work. The company was labeled a supply-chain risk and blocked from the GSA's central AI platform GuruFocus. Anthropic fought back in court and won a preliminary injunction that forced the GSA to reverse the ban.
OpenAI's discounted deal sidesteps those politics by focusing on civilian agency work. But the competition is real. Whoever controls the government's AI tooling controls a $1+ billion market and influence over how America's bureaucrats think and work. This contract is OpenAI's insurance policy.
Publishers
13
Articles
6
Reach
19