Google Vertex AI

Short Intro: Google Vertex AI, now rebranded as the Gemini Enterprise Agent Platform as of April 2026, is Google Cloud's fully managed AI platform giving teams access to 200+ foundation models including Gemini 3, Anthropic Claude, Meta Llama, and open Gemma models through a single API endpoint that remained backward compatible through the rebrand. The platform restructured its hierarchy at Cloud Next 2026, placing agents above models in the console — a deliberate architectural bet that agents are the primary unit of AI work, with model access and training now nested as sub-features beneath an agent-first interface. The Agent Development Kit, Google's primary open-source contribution to the ecosystem, reached v1.0 stable across Python, Go, Java, and TypeScript at the same event.

Key Capabilities:

• Model Garden with 200+ foundation models including Gemini 3, Claude, Llama, and Gemma

• Agent Engine managed runtime for deploying, scaling, and monitoring AI agents in production

• ADK (Agent Development Kit) open-source framework in Python, Go, Java, and TypeScript

• Agent Studio low-code visual builder for agent configuration from natural language descriptions

• Agent Engine Sessions and Memory Bank for persistent context across agent interactions

• Agent Registry, Agent Identity, and Agent Gateway for enterprise governance

• Managed MCP servers with Apigee as an API-to-agent bridge

• A2A (Agent-to-Agent) protocol v1.0 for multi-agent communication

• Vertex AI Search for natural language retrieval across internal documents

• Project Mariner web-browsing agent for computer use workflows

• Partner agents from Box, Workday, Salesforce, and ServiceNow

• Workspace Studio no-code agent builder connected to Google Workspace

• AutoML, custom model training, MLOps pipelines, and model registry under the Models sub-menu

• Vertex AI Workbench with integrated JupyterLab for data science workflows

See Google Vertex AI pricing details →