Vertex AI now offers unparalleled choice and control by allowing users to securely deploy a growing selection of leading proprietary models directly within their Virtual Private Cloud (VPC). This significant enhancement provides flexibility for building advanced AI applications. Users can access these models, including closed-source options and those with restricted licenses, through the Vertex AI Model Garden. This central hub features over 200 foundation models, encompassing Google's Gemini family, open models, and third-party offerings. Organizations requiring maximum data and infrastructure control can now self-deploy these powerful proprietary models, ensuring their sensitive business data remains within their environment and adheres to VPC Service Controls. Deployment offers optimization for performance or cost, with manual or auto-scaling options and the ability to select specific Google Cloud regions. The platform simplifies discovery, licensing, and deployment through a curated catalog and a few clicks, with a fully managed AI inference service handling the underlying infrastructure. Pricing is pay-as-you-go, allowing cost control through scaling and the application of existing Google Cloud discounts. The launch includes models from eight partners, covering diverse use cases like code generation, realistic text-to-speech, and image-to-3D model transformation. Getting started involves visiting the Model Garden, selecting "Self-deploy partner models," enabling the desired model with a license purchase, and then deploying it with a few clicks. This commitment to an open and flexible AI platform empowers responsible innovation.