Gemini 2.5 Now Live on Vertex AI: Pro, Flash & Model Optimizer

Gemini 2.5 Models Now Generally Available on Vertex AI

Google Cloud has announced the general availability of the Gemini 1.5 Pro model on Vertex AI, alongside a new lightweight model, Gemini 1.5 Flash. Both models are now accessible for enterprise developers looking to build AI-powered applications at scale.

What’s New in Gemini 1.5 Pro

Gemini 1.5 Pro excels at long-context understanding with support for up to 1 million tokens. This makes it ideal for processing large documents, codebases, or customer interactions. It also shows improved performance across reasoning, instruction-following, and coding tasks.

Introducing Gemini 1.5 Flash

Gemini 1.5 Flash is optimized for speed and efficiency. It’s ideal for high-volume, low-latency use cases like summarization, chat, and real-time data extraction. While it's smaller and faster than Pro, it retains impressive performance in key areas like summarization and question-answering.

Built on a Shared Architecture

Both 1.5 Pro and Flash are built on the same Mixture-of-Experts architecture, dynamically activating the most relevant parts of the model per task. This enables more efficient resource usage and faster inference.

Seamless Integration with Vertex AI

With Vertex AI, developers get access to robust enterprise tools like grounding with Google Search, function calling, and multimodal input support. Gemini models are also integrated with model evaluation and tuning capabilities, allowing teams to fine-tune behavior and performance easily.

Enterprise-Ready and Secure

All Gemini models on Vertex AI offer enterprise-grade security, compliance, and data governance. Users retain control over their data, and Google Cloud ensures that models do not train on customer data unless explicitly authorized.

Model Optimizer and New Evaluation Tools

Vertex AI also introduces the Model Optimizer, a new tool to fine-tune and distill foundation models for better performance in production. Additionally, automatic evaluations allow quick benchmarking and improvements based on real use cases.

Getting Started

Developers can access Gemini 1.5 Pro and Flash via the Vertex AI Studio or APIs, with flexible pricing and quota options. These models are also powering other Google services, including Workspace and Search.

To explore more or start building, visit the Vertex AI homepage.