OpenAI slashed prices on two of its AI models Thursday, days after CEO Sam Altman met with lawmakers and White House officials in Washington to discuss cybersecurity and the company’s upcoming model.
The company cut the price of its Luna model by 80%, dropping the cost of sending text to 20 cents per million tokens, down from $1, and the cost of generating responses to $1.20 per million tokens, down from $6. Terra, a mid-tier model, was cut 20%, with input costs falling to $2 per million tokens from $2.50 and output costs to $12 from $15. The price of OpenAI’s flagship Sol model was not changed.
“Our strategy remains focused on advancing both capability and efficiency so each generation of intelligence can accomplish more work at a lower cost,” OpenAI said in announcing the cuts.
Why now
Reuters reported the move comes as businesses grow more sensitive to AI costs and as US companies face pressure from cheaper Chinese rivals. OpenAI said the price drop was partly enabled by efficiency gains in the underlying GPT-5.6 model.
The cuts follow a similar move by Microsoft, whose CEO Satya Nadella highlighted the company’s cost-effective models on an earnings call this week after launching what Microsoft described as a cheap-but-performant cybersecurity model. Google also released three new lower-cost models this month.
For comparison, Anthropic’s mid-tier Claude Sonnet 4.6 model costs $3 per million input tokens and $15 per million output tokens — above OpenAI’s newly discounted Terra pricing.
The Washington visit
Altman’s pricing announcement came just after a series of meetings in Washington. He briefed lawmakers and Trump administration officials on OpenAI’s upcoming AI model and discussed cybersecurity, according to a person familiar with the plans. He told reporters Wednesday he has spoken with White House officials about the need to slow the pace of AI development.
The visit followed a security incident this month in which an OpenAI AI system gained unauthorized access to the startup Hugging Face and later to a customer account on the cloud platform Modal, using it to launch further attacks. Altman said he was supportive of congressional efforts to pass AI cybersecurity legislation.
The meetings also come as the Trump administration weighs whether to restrict access to Chinese open-weight AI models. A coalition including Nvidia, Microsoft, Meta and Palantir has urged policymakers to avoid “premature restrictions” on such models; OpenAI signed the letter after its initial publication.
