OpenAI Just Made ChatGPT Free for 23 Million Government Workers
OpenAI and the U.S. General Services Administration have signed a new multi-year OneGov agreement, running October 1, 2026 through December 31, 2028, that replaces OpenAI’s prior $1-per-agency-per-year pilot with a $0 license fee — normally $15 per user per month — plus a 50% discount on usage costs and no minimum commitment. OpenAI says the deal extends ChatGPT access to roughly 23 million federal, state, local, and tribal government workers, up from the more than one million already using it through existing programs, and bundles in discounted access to Daybreak Blue for cyber defenders, ChatGPT Enterprise data protections, and an expanded Government Academy. OpenAI is citing early results to justify the push: the CDC cut literature reviews from days to under 30 minutes, Georgia reduced tax-form digitization from two weeks to 15 minutes, and Lawrence Livermore National Laboratory compressed fusion-research development from months to hours.
OpenAI isn’t alone in treating government as a strategic wedge. Google DeepMind’s Fairwind Program, announced a week earlier, gives government agencies and critical-infrastructure operators early access to its Gemini 3.8 Flash Cyber model paired with automated vulnerability-patching technology, and already counts more than 650 participating partners including CrowdStrike and Palo Alto Networks; Google says it has committed over $100 million to cybersecurity initiatives to support it. Anthropic made its own government play in July, bringing Claude Code and Claude Cowork into a FedRAMP High authorized environment for Claude for Government Desktop, with agency-level spend controls and tamper-evident audit logs built for the federal Authority to Operate process.
Three frontier labs, three government-specific packages, inside about ten weeks of each other. Government isn’t a side market these companies are getting around to — it’s now a competitive line item each one is racing to lock down, on price, on compliance certification, or on both at once.