Google just launched a big upgrade to Android Studio. The new feature? Agent Mode, an experimental AI tool powered by Gemini that lets developers handle complex coding tasks through simple chat commands.
This isn’t your average autocomplete. Agent Mode lets devs describe goals in plain language — like generating unit tests or refactoring code — and then builds and executes a multi-file plan. It’s all controllable, stopping at checkpoints so devs can tweak or reject outputs. For those who want speed, an “auto-approve” option speeds things through.
This build rides on the Gemini 2.5 Pro model, which offers a massive 1 million token context window. That’s huge for managing enormous codebases and complex workflows. The basic Gemini model is free with daily quotas, but paid subscribers unlock this bigger context window via API keys. 
Agent Mode taps into the Model Context Protocol, an open standard created by Anthropic, allowing it to interact with external tools and other AI agents. Collaborative coding just got a major upgrade.
Google showed off this tech at Google I/O 2025. Now, devs can test it today by downloading the Android Studio preview on the Narwhal canary channel.
Google explained the benefits in their blog:
“You can delegate routine, time-consuming work to the agent, freeing up your time for more creative, high-value work.”
Try the new Agent Mode now and see how AI runs your code. The future of Android app building just got a lot faster.
