AI Lab Logo
Open and closed models · Gemini, Gemma, Imagen, Veo

Google models in Australia

Google DeepMind develops the Gemini family of multimodal models, which accept text, images, audio, video and documents, and power the Gemini app, Google Workspace features and developer APIs. Gemini is offered in Pro, Flash and Flash-Lite tiers, alongside Imagen for images and Veo for video.

Google also releases Gemma, a family of open-weight models that can be run on laptops, servers or cloud platforms. Gemma models can be downloaded for local use, and some are available in-region on Amazon Bedrock in Sydney and Melbourne, which gives Google-built models an onshore route even outside Google Cloud.

For production use with data controls, developers use Gemini through Google Cloud's Gemini Enterprise Agent Platform, formerly Vertex AI. There, selected Gemini models support machine learning processing in the Sydney region when called through a regional endpoint, while Google AI Studio and the consumer app suit prototyping and personal productivity.

Last reviewed October 2026

Strengths

  • +Native multimodal input, including long video, audio and large document collections
  • +Very long context windows for analysing large codebases or document sets
  • +Selected Gemini models process data in Google Cloud's Sydney region
  • +Tight integration with Google Workspace, Search grounding and Google Cloud data services
  • +Open-weight Gemma models for local or self-hosted use

Limitations

  • –Only some Gemini models, mostly Flash tiers, are listed for processing in Australia
  • –Newest Gemini models usually launch on global endpoints with no residency commitment
  • –Google AI Studio and the consumer Gemini app are not designed for Australian data residency
  • –Product naming changes often, which makes documentation harder to track

Ways to access these models

Gemini app and Google Workspace

Gemini assistant features for business users inside Gmail, Docs, Sheets and Meet.

Google AI Studio and Gemini API

Quick prototyping and API keys for developers, without regional processing controls.

Google Cloud Gemini Enterprise Agent Platform

Formerly Vertex AI; regional endpoints, enterprise controls and partner models.

Self-hosted or Bedrock Gemma

Run open Gemma weights locally or use them in-region on Amazon Bedrock in Sydney.

Keeping data in Australia

Google Cloud's data residency documentation lists selected Gemini Flash models and a text embedding model as supporting machine learning processing in Australia (australia-southeast1, Sydney) when you call the regional endpoint. Requests to the global endpoint may be processed in any Google Cloud location and carry no residency commitment, so restrict access to regional endpoints with organisation policies. Partner and open models on Google Cloud, including Claude, Llama, Mistral and Qwen, are not listed for Australia. Gemini Enterprise offers an Australian in-country multi-region on an allowlist basis through your Google account team. Gemma can be self-hosted or used in-region on Amazon Bedrock in Sydney and Melbourne. Confirm against Google Cloud's data residency table for the exact model version before deployment.

For workloads where data must stay onshore end to end, see private AI hosted in Australia.

Best for

  • →Google Cloud customers who need Gemini processed in Sydney
  • →Multimodal analysis of video, audio, images and long documents
  • →Organisations standardised on Google Workspace
  • →Local or self-hosted workloads suited to Gemma

Tools built on Google models

  • Google Gemini

    Free plan + paid

    General Assistant

    Google's AI assistant, built into Google Workspace. It works inside Gmail, Docs, Sheets and Drive, can search the web, and understands text, images, audio and video.

    Best for: Businesses already running email and documents on Google Workspace

    Google Gemini review →

  • Google AI Studio

    Free plan + paid

    Dev Tools

    Google's free browser-based workspace for trying Gemini models, testing prompts, generating API keys and building apps. Its Build mode generates complete web or Android apps from a prompt, which you can export to GitHub or deploy to Google Cloud Run.

    Best for: Developers prototyping with Gemini before production

    Google AI Studio review →

  • Gemini Notebook (formerly NotebookLM)

    Free plan + paid

    Research/Study

    Google's AI research notebook, renamed Gemini Notebook in July 2026 (formerly NotebookLM). You add documents, and it answers questions, writes reports and study guides, and produces audio and video overviews, mind maps and slide decks, all grounded in the sources you provided.

    Best for: Consultants and analysts working through long reports

    Gemini Notebook review →

  • Google Veo

    Paid

    Video Generation

    Google DeepMind's video generation model family. It creates short, realistic video clips from text or image prompts, keeps characters consistent using reference images, extends scenes, and can generate matching sound effects, music and dialogue. It is used in the Gemini app and Google Flow.

    Best for: Marketing teams prototyping video ads before committing to a shoot

    Google Veo review →

  • Ollama

    Free plan + paid

    Dev Tools

    An open-source tool for downloading and running open-weight language models on a Mac, Windows or Linux machine, or on your own server. It provides a simple command line, a desktop app and a local API, plus optional paid cloud models for larger workloads.

    Best for: Developers prototyping with open models on their own machines

    Ollama review →

Frequently asked questions

Which Gemini models can run in Australia?

Google Cloud lists selected Gemini Flash models, plus a text embedding model, for machine learning processing in australia-southeast1 (Sydney). The newest Pro models are often available only on global endpoints at launch. Check Google's data residency table for the specific model version, because coverage changes as models are added and retired.

Is Vertex AI still the right place to use Gemini for business?

Yes, although Google now calls the developer platform the Gemini Enterprise Agent Platform, formerly Vertex AI. It provides regional endpoints, identity and access controls, logging and enterprise terms, which Google AI Studio's API keys do not. Use regional endpoints if you need Australian processing.

Does the free Gemini API keep data in Australia?

No. The Gemini API through Google AI Studio has no Australian processing option, and free tiers can allow Google to use content to improve its products. For business data, use a paid Google Cloud project with regional endpoints, or Gemini features covered by your Google Workspace agreement.

What is Gemma and how is it different from Gemini?

Gemma is Google's family of open-weight models built from the same research as Gemini. You can download Gemma and run it on your own hardware, which gives full control over where data is processed, but it is smaller and generally less capable than Google's hosted Gemini models.

Other model providers

Building on Google models?

AI Lab Australia designs and builds AI systems on the right model and hosting route for your privacy and quality requirements.

Talk to AI Lab AustraliaBrowse all AI tools