OpenAI has launched a comprehensive economic framework to measure artificial intelligence return on investment, shifting corporate evaluation from traditional adoption metrics to a standard termed Useful Intelligence per Dollar. Introduced alongside the GPT-5.6 release, the initiative provides executives with a structured approach to quantify AI expenditure against tangible operational output. The framework replaces legacy indicators like cost per token or software licenses, emphasizing the full economic cost of delivering a successful outcome. OpenAI outlines four foundational pillars: quantifying completed work, calculating cost per successful task, measuring system dependability, and tracking economic improvement at scale. By focusing on finished workflows rather than raw model interactions, organizations can align AI deployment with strategic objectives. Support teams can track resolved customer issues, engineering divisions can monitor deployed code, and legal departments can measure contract review efficiency. Central to the rollout is GPT-5.6, a new model family featuring three tiers to optimize performance and expenditure. The flagship Sol tier delivers advanced reasoning, achieving a new state of the art on the Artificial Analysis Coding Agent Index while utilizing 54 percent fewer output tokens than competing systems. The Terra tier balances depth and cost, while the Luna tier provides rapid, high-volume processing. OpenAI advises that model selection should be dictated by workflow complexity and first-pass success rates rather than base inference pricing. Reliability remains critical to the evaluation model. As AI systems transition from drafting assistance to executing complex workflows, dependability directly influences operational economics. Accurate, well-sourced outputs reduce human review cycles, lower total compute requirements, and foster enterprise confidence. To support this transition, ChatGPT Work extends the security, privacy, and compliance architecture of ChatGPT Enterprise, enabling deeper system integration while maintaining strict oversight boundaries. The company also highlights the compounding economic benefits of infrastructure scaling. Improved training compute, purpose-built hardware, optimized routing, and algorithmic efficiency collectively enhance model capability and inference speed. These advancements distribute across OpenAI’s ecosystem, including the ChatGPT platform, enterprise workspaces, developer tools, and the API. OpenAI projects that as each platform layer improves, customers will experience faster results, fewer corrections, and declining costs per task. The Useful Intelligence per Dollar framework positions AI as a direct driver of operational efficiency rather than an experimental technology. By prioritizing measurable work output, economic transparency, and dependable execution, OpenAI aims to accelerate enterprise adoption and redirect human labor toward high-value judgment and creative tasks.