Google is opening its Android Studio development environment to more than 40 external coding agents like Claude and Codex via the Bring Your Own Agent feature, upending the single-vendor model that has historically defined integrated development environments.
As Google introduced updates for its developer tools and system APIs at this year's Android Dev Summit, the decision to dismantle its proprietary moat marks a structural shift in how software engineering toolchains are monetized and maintained. Can external agents truly operate efficiently within Android Studio without exploding inference costs and token consumption?
With Android Studio Rabbit 2, Google is launching a Bring Your Own Agent feature. Developers are no longer limited to Google's own AI tools. Instead, they can use the open Agent Client Protocol to run more than 40 different coding agents directly in the IDE. Developers can install and simultaneously run Anthropic's Claude Agent, OpenAI's Codex, and Google Antigravity through the built-in marketplace, granting these external systems direct access to project structure, symbol resolution, the emulator, and the Logcat system.
Architectural Shifts and Token Economics
The integration of multi-vendor orchestration forces a fundamental recalculation of compute budgets for engineering teams. To measure model performance on complex tasks, Google released version 2.0 of the Android Bench leaderboard, which evaluates language models on multi-day development projects based on functionality, time spent, and cost.
Internal Google tests show agents finished UI tasks 77 percent faster and used up to three times fewer tokens when rendering Compose previews in the terminal.
This performance gain directly impacts the unit economics of AI-assisted engineering by curbing the runaway token expenditure typical of round-trip cloud API calls. Furthermore, Google is testing Lightbuild, a declarative build system based on a YAML schema that both humans and AI agents can read and edit, alongside a code verification tool called Change Review that analyzes pending changes, generates bug reports, and explains modifications step by step.
Local Execution and Performance Tooling
The expansion of command-line capabilities moves processing closer to the physical device layer. Google expanded the Android CLI command-line tool for developers who run their agents in the terminal or in CI pipelines, making it available as an official agent plugin for Claude Code, OpenAI's Codex, and Google Antigravity. In a Google A/B test using Anthropic's Claude Opus 5.5 model, the CLI tool helped find and fix memory leaks nearly twice as fast while cutting token usage by about 25 percent because Google built the open-source LeakCanary memory leak detection library into both the CLI and the Android Studio Profiler to evaluate heap dumps on the computer instead of the mobile device.
Beyond desktop environments, Google is shifting execution to mobile hardware through local architecture. For local on-device processing, Google is adding an alpha Agentic API powered by Gemini Nano 4 to its ML Kit.
Google is using the Model Context Protocol and launching the AppFunctions API in beta as a unified tool-calling layer so apps can share functions with local or system-wide agents. As Google and hardware maker XREAL also unveiled the XREAL AURA, a $1,099 pair of smart glasses running Android XR that displays virtual workspaces and uses Gemini to analyze the real-world environment.
A multi-vendor IDE architecture running over forty external agents inside Android Studio relies entirely on standardized protocols remaining open while inference costs scale downward. Regulatory pressure, demonstrated when on 2026-07-17, as we noted in our earlier story EU Forces Google to Open Android to AI Rivals by 2027, forced the platform to yield its walled garden, leaving margins to be squeezed between open APIs and the cost of maintaining local ML Kit runtimes.
Against the field
While the new Android Bench 2.0 leaderboard pits Claude against Codex on speed and token spend, Google ensures that keeping engineers inside its build tools and CLI matters far more than which model writes the code. Google turns an inevitable antitrust concession into a strategy for maintaining ownership of the developer environment.
Our stories: EU Forces Google to Open Android to AI Rivals by 2027 (July 17).