TapTech News, May 14 - OpenAI released its latest flagship artificial intelligence model GPT-4o early this morning. Compared to previous versions, GPT-4o has seen many upgrades: speed doubled, price lowered by 50% (processing cost of one million tokens reduced to $7), and supports processing up to tens of millions of tokens per minute. The model has a memory capacity of 128K contexts, with knowledge cut-off date of October 2023. Microsoft announced that a preview version of GPT-4o is available through Azure OpenAI service.
Unlike previous models, GPT-4o can seamlessly process text, images, and audio information, providing users with a richer and more immersive interactive experience. However, TapTech News noted that in the initial version, GPT-4o only supports text and images, with audio and video support to be added in the future.
Azure OpenAI users in some parts of the United States can be the first to experience the powerful capabilities of GPT-4o through the preview feature of Azure OpenAI Studio. Microsoft will share more updates about GPT-4o and Azure AI at the Build 2024 conference later this month.
Related Reading:
OpenAI Releases New Flagship Generative AI Model GPT-4o: Smoother Voice Conversations, Free of Charge