OpenAI Cracks Code on Cheaper AI Without Sacrificing Quality
OpenAI has set its sights on delivering high-quality artificial intelligence (AI) while reducing costs, according to CEO Sam Altman. The strategy involves distilling larger models into smaller, cheaper versions that maintain competitive performance.
The company's efforts are already yielding results. In July 2026, OpenAI cut API pricing for its GPT-5.6 family of models, slashing the Luna model's costs by 80% and Terra's by 20%. This price reduction led to a surge in Luna consumption, with usage growing roughly 14-fold.
The effective price dropped by about 10-fold, resulting in a 34% increase in Luna revenue. OpenAI attributed these gains to efficiency improvements, including a 54% reduction in output tokens and over 20% decrease in serving costs.
In September 2026, the company announced further price cuts for its newer GPT-6 model variants, pricing them roughly half what the previous series cost. This move came as other models like Kimi K3 began applying pressure on the cost-performance frontier.