OpenAI introduced GPT-5.6 Luna and Terra with significant price reductions and performance improvements. Luna, the fastest and most affordable model, will cost 80% less, while Terra, a balanced model for everyday work, will cost 20% less. These updates aim to enhance cost-effectiveness and speed for enterprise workloads. The changes reflect years of improvements in model efficiency and infrastructure, aligning with OpenAI's mission to make advanced intelligence more accessible. Source: openai
GPT-5.6 Luna delivers performance comparable to frontier-class models from a year ago at roughly 6 cents on the dollar per task, nearly nine times faster. On professional work, as measured by Agents’ Last Exam, Luna outperforms Fable 5 at an estimated cost per task nearly 99% lower. Businesses can define the outcome and quality standard they need, then use evaluations to determine where additional intelligence improves results and where faster, lower-cost processing can deliver the same quality. A coding workflow, for example, might use Sol to resolve uncertainty and define the plan, then use Luna to implement well-specified changes, write and run tests, and evaluate the results. Another workflow may call for a different balance. Source: openai
OpenAI’s efficiency gains come from improving models, inference systems, and the agentic harness connecting them to tools and context. GPT-5.6 models take a more direct path through work, with better routing keeping hardware productive, optimized production software generating tokens more efficiently, and smarter context management helping agents avoid repeating completed work. These improvements allow completing more useful work with the same compute, reducing time, tokens, and cost per result. Source: openai
