Microsoft has announced the availability of OpenAI's latest inference models, o3 and o4-mini, on its Azure OpenAI services within the Azure AI Foundry. These models offer substantial performance enhancements over their predecessors, o1 and o3-mini. The o3 model showcases groundbreaking capabilities in various domains, including coding, mathematics, science, and visual perception, particularly excelling in visual tasks such as image processing. Conversely, the o4-mini model is a compact version designed for rapid and cost-effective inference, delivering impressive performance in mathematics, coding, and visual tasks.
