OpenAI has reportedly implemented significant optimizations to its AI models, resulting in a reduction of inference costs by over 50%. According to a recent report by The Information, these changes have particularly benefited the ChatGPT platform, where the number of Nvidia GPUs required for operations has been reduced to just a few hundred at times. This strategic move not only streamlines the operational costs for OpenAI but also enhances the service's accessibility for guest users, potentially attracting a larger user base. The implications of these cost reductions could foster increased usage and engagement, positioning OpenAI favorably in the competitive AI landscape.
Source: The Decoder
