Google Cloud announced updates and enhancements to its generative AI models and platform, Gemini and Vertex AI, to help businesses move from experimentation to production.
One key area addressed was how to maximize model performance without compromising on latency and costs. The Gemini 1.5 Pro and Flash models have been updated to deliver quality improvements in math, long context understanding, and vision. Additionally, Gemini 1.5 Pro pricing has been reduced by 50% across both input and output tokens.
Imagen 3, an image generation model, has also been enhanced with improved prompt understanding, instruction following, and photorealistic quality. Other updates include the ability to quickly insert or remove objects and make background changes, as well as powerful new customization features.
To address reliability concerns, Google Cloud introduced new capabilities in Vertex AI, including Controlled Generation, which allows for fine-grained control over AI outputs, and Prompt Optimizer, which assists in improving model performance.
Furthermore, Google Cloud reinforced its commitment to a “your AI, your way” approach by offering flexible AI deployment options, including data residency options during ML processing.
Overall, these updates aim to equip businesses with the tools and technologies needed to deploy generative AI solutions at scale with confidence.