OpenAI may be about to make using its AI a lot cheaper—and not out of charity. The Wall Street Journal, citing "people familiar with the matter," reports the company is weighing steep cuts to the prices it charges per "token," the basic unit that underpins AI billing, in anticipation that rival Anthropic will do the same. OpenAI currently has tiered subscriptions of $8, $20, and $100 or more per month for its GPT 5.5 models, CNBC reports, while Anthropic users pay $17 a month for Claude Pro and $100 or more for Claude Max.
CEO Sam Altman recently acknowledged that usage costs have become "a huge issue" for corporate customers, even as both firms already burn through billions on computing power. "I think we'll have a lot of ways we can help people get more value for less spend," he said. A price war could squeeze margins just as the two leaders in generative AI prepare to face public markets. Both companies have been losing billions of dollars a year but Anthropic's revenue has recently surged, boosted by the popularity of its Claude Code tool with software engineers and a valuation that has now topped OpenAI's, the Journal reports.
That momentum, along with executives' growing unease over "tokenmaxxing"—pouring money into AI usage without clear returns—is pushing OpenAI to sharpen its pitch and its pricing as it races for enterprise customers and readies an IPO targeted within a year. Anthropic filed for an IPO a week earlier and it's not clear which company will win the race to Wall Street.