Apple integrates Google Gemini into Xcode 27, expanding native agentic coding options
Apple's Xcode 27 adds native support for Google Gemini, joining existing integrations with Anthropic's Claude and OpenAI's Codex. The update also introduces improved interfaces, interactive planning, and multiturn Q&A capabilities for AI-assisted development.
Apple integrates Google Gemini into Xcode 27, expanding native agentic coding options
Apple's Xcode 27 now supports Google Gemini natively for agentic coding tasks, joining existing integrations with Anthropic's Claude and OpenAI's Codex. The update also adds new interfaces, workspaces, interactive planning, and multiturn Q&A capabilities.
Native Gemini integration
Developers can now configure Gemini directly through Xcode's Intelligence settings panel without needing external plugins. According to Google, the integration enables "complex, multi-step tasks during development without switching tools or windows," including code review, bug fixes, and feature development.
This marks the third natively supported AI assistant in Xcode, following the initial agentic coding rollout in Xcode 26.3 earlier this year. That release introduced direct integration with OpenAI Codex and Anthropic's Claude Agent, plus an "Add an Agent" tool for custom AI connections and MCP (Model Context Protocol) support for external agents.
Expanded agentic capabilities
Xcode 27 builds on the agentic framework established in version 26.3, which allowed AI tools to handle tasks from writing boilerplate code to updating entire projects based on documentation and file structure. Version 26.5 added message queueing and the ability for agents to ask clarifying questions.
The new release adds:
- New interfaces and workspaces for agent interactions
- Interactive planning features
- Multiturn Q&A support for more complex conversations with AI assistants
Apple has not disclosed pricing for the Gemini integration or whether it requires separate Google Cloud credentials. The company also has not specified which Gemini model version (1.5 Pro, 2.0 Flash, or others) is supported.
What this means
Apple is treating IDE-integrated AI coding as a multi-vendor feature rather than building a proprietary solution, following the pattern Microsoft established with GitHub Copilot's multi-model support. By adding Gemini alongside Claude and Codex, Apple gives developers flexibility to choose AI assistants based on task requirements or existing subscriptions. The multiturn Q&A and interactive planning features suggest Apple is moving toward more sophisticated agentic workflows where AI tools can handle sustained, multi-step development tasks rather than single code completions. The MCP support means developers aren't locked into the three native options—any compatible AI agent can plug into Xcode's framework.
Related Articles
GitHub Copilot App Adds Guide for Running Multiple Agents in Parallel
GitHub has published a beginner-focused guide explaining how to run multiple agents simultaneously in the GitHub Copilot app. The guide targets developers unfamiliar with parallel agent workflows, framing the shift from single-agent to multi-agent use as a key productivity unlock.
GitHub Publishes Beginner Guide to Running Multiple Copilot Agents in Parallel
GitHub has published a beginner-focused guide explaining how to run multiple Copilot agents concurrently within the GitHub Copilot app. The post walks new users through the workflow of parallel agent execution rather than announcing a new model or feature.
Anthropic Brings Background Computer Use to Claude Code and Cowork on Mac
Anthropic has enabled background computer use for Claude Code and Claude Cowork on macOS, available to Pro and Max subscribers. The feature lets Claude click, type, and open apps on a Mac without taking over the user's active cursor, following a similar launch by OpenAI's ChatGPT earlier in 2026.
GitHub Explains How Copilot Cuts AI Coding Costs Without Lowering Task Quality
GitHub published an engineering breakdown of how Copilot reduces cost per coding task by targeting wasted work across the full task lifecycle, not just shortening model outputs. The post argues that shorter responses alone don't guarantee lower cost or better efficiency.
Comments
Loading...