Use Claude on Vertex AI with multimodal, tool use, and prompt caching

We added a new Claude model integration backed by Vertex AI, enabling you to run Claude where your data and governance live on Google Cloud. This unlocks multimodal inputs, tool calling, prompt caching, and advanced “thinking” responses — without changing your application code. Teams can standardize on Vertex AI while keeping model choice flexible and costs predictable.

Details

Multimodal support for image and text inputs
Native tool use to orchestrate actions and APIs
Prompt caching to lower latency and reduce spend
“Thinking” mode for more reliable, stepwise reasoning

Who this is for: Teams standardized on GCP/Vertex AI, and builders needing multimodal and tool-rich AI with stronger governance.