Features

Access to Foundation Models and APIs

Choose the right model for your use case with 40+ proprietary models and 60+ OSS and 3rd party models on Vertex AI's Model Garden. With access to Google's foundation models as APIs, you can easily deploy these models to applications.

Choose from Google Research like PaLM 2, Imagen, and Codey, and other models like Llama 2 and Claude 2.

Experiment and test models with prompt design

Adapt models to your use case with prompt design. Iterate on the prompt through a familiar chat interface and choose from multiple ways to adjust responses. For example, you can change the response "temperature" to elicit a more creative response. 

Easily tune the model with your own data

Improve the quality of model responses for your use case by tuning foundation models with your own data with Vertex AI's Generative AI Studio.

Access state-of-the-art tuning options like adapter tuning and Reinforcement Learning from Human Feedback (RLHF) or style and subject tuning for image generation. 


Connect models to real-world data and real-time actions

Vertex AI Extensions provides a set of fully-managed tools for building and managing extensions that connect models to proprietary data sources or 3rd party services. Now developers can create generative AI applications that deliver real-time information, incorporate company data, and take action on the user's behalf. 

Integration with end-to-end ML tools

Vertex AI's managed endpoints make it easy to build generative capabilities into an application, with only a few lines of code and no ML background required. Developers can forget about the complexities of provisioning storage and compute resources, or of optimizing the model for inference. 

Once deployed, foundation models can be scaled, managed, and governed in production using Vertex AI’s end-to-end MLOps capabilities and fully managed AI infrastructure. 


Enterprise-grade data governance and security

With Vertex AI, your data is completely protected, secure, and private when using it to customize a model, and you have full control over where and how or if their data is used. None of the customer’s data, model weights, or input prompts are used to tune the original foundation models. When enterprises tune a model with their own data, the original model remains unchanged, and the new model never leaves your company’s environment. 

How It Works

Get started quickly foundation model APIs

Learn to tune foundation models
Learn how to tune foundation models

Common Uses

PaLM 2: Google's next gen LLM

Use PaLM 2 for a wide range of tasks

PaLM 2 excels at a variety of capabilities with improved multilingual, reasoning, and coding capabilities. From generating and translating nuanced text, including idioms, poems, and riddles to demonstrating logic, common sense reasoning, and mathematics, PaLM 2 is more capable, faster, and more efficient than previous models.
Explore sample prompts
PaLM 2

Use PaLM 2 for a wide range of tasks

PaLM 2 excels at a variety of capabilities with improved multilingual, reasoning, and coding capabilities. From generating and translating nuanced text, including idioms, poems, and riddles to demonstrating logic, common sense reasoning, and mathematics, PaLM 2 is more capable, faster, and more efficient than previous models.
Explore sample prompts
PaLM 2

Imagen: Generate and customize images

Studio-grade images at scale for any business need

With Imagen, it’s easy to create and edit high-quality images at scale with low latency and enterprise-grade data governance. Organizations can also customize and adapt Imagen to their business with object tuning and style tuning Leveraging the power of mask-free edit, image upscaling, and image captioning across over 300 languages, customers can quickly generate production ready images.
Watch a video to learn more about Imagen
Imagen thumbnail