Altman: Token-Based Pricing Becoming Obsolete as GPT-5.5 Shifts Focus to Task Completion Over Token Count

Gate News message, April 29 — OpenAI CEO Sam Altman said in an interview with Ben Thompson on Stratechery that token-based pricing is not a long-term viable model for AI services. Using GPT-5.5 as an example, Altman noted that while the per-token price is significantly higher than GPT-5.4, the model uses far fewer tokens to complete the same task, meaning customers do not care about token count—they only care about whether the task is completed and the total cost.

"We are not a token factory; we are more like an intelligence factory," Altman said. "Customers want to buy the most intelligence for the least money. Whether the underlying work is done by a large model running few tokens or a small model running many tokens does not matter to them." He added that OpenAI's current customer base is increasingly demanding more capacity rather than negotiating prices, with far more customers saying "give us more capacity, no matter the cost" than those asking for discounts.

Drawing a parallel to utilities, Altman explained that unlike water or electricity—where lower prices do not significantly increase consumption—AI demand scales differently. "As long as the price is low enough, I will keep using more. No other public utility works this way," he said. AWS CEO Matt Garman added that computing power prices have dropped by multiple orders of magnitude over the past 30 years, yet more compute is being sold today than ever before.

Altman also characterized ChatGPT as "the first truly large-scale consumer product since Facebook," acknowledging that while AI was expected to disrupt search, the real wins came from ChatGPT itself and the Codex API. He noted that "Google is still underestimated in many ways."