OpenAI reportedly discovered a new optimization method that could reduce inference costs by over 50%.

Odaily Planet Daily News According to The Information, OpenAI's engineering team recently told some colleagues that the company has found a new system optimization method that can reduce the "inference" cost of AI models by more than half. Inference cost refers to the computing resource cost consumed when the model is actually running and responding to user requests. This optimization mainly comes from improving the utilization efficiency of existing server resources, rather than relying on new computing chips. This progress reflects that while AI companies continue to compete for computing resources, they are also improving the efficiency of existing infrastructure through software and system-level optimization to alleviate the rapidly growing cost pressure of model operation.