Microsoft has made its GPT-4 Turbo with Vision model generally available for its Azure OpenAI Service customers. This model was initially launched in December 2023 and is now available for deployment in the Sweden Central and East US 2 Azure OpenAI regions. The model has been used by customers and partners to enhance efficiencies and innovate within their businesses, with applications ranging from improving online shopping experiences to deriving insights from charts and diagrams.
However, some features like Optical Character Recognition (OCR) and object grounding are not available in the general release version, but Microsoft is exploring ways to integrate them in future updates. The company also plans to add a JSON mode and function calling for inference requests involving image inputs in the near future.
Microsoft has also announced pricing for the GPT-4 Turbo with Vision model, with rates of $0.01 per 1000 tokens for Input and $0.03 per 1000 tokens for Output. Different pricing will apply for enhanced features.
Additionally, Microsoft is testing a new “Risks & safety monitoring” feature in Azure OpenAI Service to provide businesses with real-time insights into how their AI systems are operating and to detect any abusive requests. This feature aims to enhance the security and monitoring capabilities of AI systems within the Azure OpenAI Service platform.
Source link
Source link: https://www.neowin.net/amp/microsoft-says-gpt-4-turbo-with-vision-is-generally-available-on-azure-openai-service/
GIPHY App Key not set. Please check settings