Google's Android Studio Quail 2 Lets Gemini Run Multiple AI Agents at Once

Android Studio Quail 2 just landed in the stable channel, and the headliner is something deceptively simple: you no longer have to wait for one AI agent task to finish before starting another. That single change reshapes how you interact with Gemini inside the IDE, and it comes bundled with two other substantial upgrades to memory leak detection and crash analysis.

One agent at a time is over

Google rebuilt Agent Mode from the ground up in Quail 2. The new architecture delivers better performance and more flexibility for decomposing complex tasks, and it's what makes running multiple agent chats simultaneously possible.

The previous bottleneck, where agent tasks could only be processed sequentially, is gone. You can now start a UI refactoring in one tab, redefine a ProGuard rule in a second, and generate documentation in a third -- all running concurrently. Each tab can even be assigned a different model.

To use it:

• Click the + icon in the Agent pane to open a new parallel conversation thread
• Use the History icon to switch between active tasks and monitor their progress
• Or go to File > New > New Agent Tab to open a conversation in a dedicated tab

One caveat: worktree support is not available with parallel agent chats. If two concurrent agents are editing the same files, you can run into conflicts, so keep that in mind when splitting up work.

LeakCanary, now a first-class citizen

Memory leaks are one of the more painful debugging experiences on Android. The object reference outlives its lifecycle, the garbage collector can't reclaim the memory, and eventually you're staring at an OutOfMemoryError. Hunting them down has historically meant a lot of manual heap analysis on a device that's already under pressure.

Starting with Quail 2, the popular open-source leak detector LeakCanary is natively integrated directly into the Profiler as a first-class task. The key trick: heap analysis is lifted off your resource-constrained test phone and moved onto your development machine. That makes leak tracing up to five times faster and jank-free, leaving the test app running smoothly on the device.

Once a leak is detected, the workflow is tight:

• The Profiler renders an interactive, color-coded leak trace, grouping occurrences and estimating lost memory.
• Click Go to declaration on any leaking object to jump directly to that line of code in your editor.
• Click Fix with Agent to have the Gemini agent ingest the trace, explain the root cause of the retained reference, and write the exact code change -- such as unbinding a listener or clearing a static reference -- to plug the leak.

Crash analysis that actually closes the loop

The third major addition is a tighter integration between App Quality Insights (AQI) -- Android Studio's panel for viewing production crash reports -- and Agent Mode. Previously, you'd look at a stack trace and then manually switch context to figure out what went wrong. Now the agent does the synthesis for you.

When you click on a crash in the AQI panel, you get an immediate high-level summary. Clicking See more opens a dedicated agent chat that pulls in your local source code alongside the full stack trace to deliver a comprehensive explanation of the failure. From there, clicking Fix with AI triggers the agent to propose a step-by-step fix plan and, on your approval, apply the changes directly to the project.

Model flexibility and pricing

Android Studio lets you bring any model into the IDE for agentic assistance -- Gemini, GPT, Claude, or a local model like Gemma 4. This flexibility gives you control over performance, privacy, and cost. You can also mix models across parallel tabs, using a heavier model for a complex refactor and a lighter one for docs generation.

If you're on a Google AI Pro or Ultra plan, you get access to dedicated capacity and higher rate limits for Gemini in Android Studio -- particularly useful for long agentic sessions that burn through tokens. Android Studio detects your subscription automatically when you log in with your Google account. For teams, enterprise-grade privacy and security features are available through Gemini Code Assist , starting at $45/user/month.

Android Studio Quail 2 is available for Windows, macOS, Linux, and ChromeOS. You can update from inside Android Studio via Help > Check for Updates, or download directly from the Android Studio site. The base IDE remains free. Notably, Android Studio Quail 3 is already available as a release candidate , so the cadence here is fast.

The parallel agent architecture is the real shift. It moves Android Studio closer to a model where the IDE is running background AI work continuously, not just responding to one prompt at a time. Combined with the LeakCanary and AQI integrations, Quail 2 makes a credible case that the most time-consuming parts of Android debugging -- not just code generation -- are now within reach of agentic automation.