OpenAI: Tests Success-Only Enterprise Pricing

OpenAI is testing outcome-based pricing for enterprise customers, billing only when AI agents complete tasks successfully while absorbing compute costs on failures, a move signaling OpenAI revenue growth is not accelerating fast enough ahead of its IPO at an $852 billion valuation. Independent tests show the Operator agent failing 62% of desktop tasks, with broader industry surveys indicating 75% completion rates, prompting clients like Thomson Reuters to build internal models and demand accountability that shifts risk back to OpenAI.