Google Cloud pushes I/O AI updates into Gemini Enterprise and Workspace

The updates bring Gemini 3.5 Flash, Gemini Omni, Spark, Antigravity, Managed Agents API, Workspace tools, and CodeMender into Google’s enterprise AI stack.

Google Cloud has announced new AI updates from Google I/O, including Gemini 3.5 Flash, Gemini Omni, Gemini Spark, Antigravity, Workspace AI tools, Managed Agents API, and CodeMender.

Google Cloud has set out a new wave of AI updates for business customers following Google I/O, moving new Gemini models, agent tools, Workspace features, and security capabilities into its enterprise platforms.

The announcements include Gemini 3.5 Flash, Gemini Omni, expanded Google Antigravity access, Gemini Spark, Google Pics, new voice features in Workspace, Managed Agents API, CodeMender, and an AI Content Detection API.

The rollout gives Google Cloud customers a clearer view of how Google is packaging its latest AI models for enterprise use: model access through Gemini Enterprise, agent development through Agent Platform, productivity features through Workspace, and security controls for code and synthetic media.

Gemini 3.5 Flash reaches enterprise users

Google is starting the Gemini 3.5 series with Gemini 3.5 Flash, which is rolling out today for developers and business users.

Developers can use Gemini 3.5 Flash to build agents through the Gemini Enterprise Agent Platform, Google AI Studio, and Google Antigravity. Business users can access the model through the Gemini Enterprise app.

Google said Gemini 3.5 Flash has been built for agentic tasks and coding, with performance focused on longer, more complex workflows. The company said Gemini 3.5 Pro is in testing and is due next month.

Google also introduced Gemini Omni, a model designed to generate video from text, audio, image, and video inputs. Gemini Omni Flash is due to roll out in the coming weeks for developers and enterprise customers through the Gemini API and Agent Platform API.

The company is positioning Gemini Omni for enterprise content work, including visual editing, video production, virtual try-ons, post-production workflows, and customer-facing media.

Antigravity and Spark target daily work

Google is extending Antigravity, its agentic development platform, into Agent Platform, giving Google Cloud customers access inside their existing cloud environment.

Antigravity 2.0 is launching as a standalone desktop app for builders to steer, customize, and orchestrate agents. Google is also adding an Antigravity CLI for developers who want a lighter interface for building and deploying agents.

The company said Antigravity will be available in Gemini Enterprise in the coming months.

Google also announced Gemini Spark, a personal AI agent for Gemini Enterprise that can work in the background across Workspace, custom connectors, and the open web.

Spark is designed to carry out multi-step tasks under user direction. It can run recurring tasks, learn skills, connect to enterprise systems, and ask for approval before high-risk actions such as sending emails.

The agent can use Gemini Enterprise connectors including Microsoft SharePoint, OneDrive, ServiceNow, and other business systems. Google said Spark runs in a managed Google Cloud runtime, with tasks executed in isolated virtual machines and traffic routed through an Agent Gateway that enforces Data Loss Prevention policies.

Gemini Spark in the Gemini Enterprise app is rolling out to customers soon. Spark in Google Workspace will be available soon in preview for business customers in the Gemini app.

Workspace and security tools get AI additions

Google is adding AI features directly into Workspace, with previews for business customers expected this summer.

Google Pics, a new AI-powered image generation and editing tool, will be built into apps including Drive, Docs, and Slides. It will allow users to move, resize, transform, and edit individual objects in images, as well as modify and translate text independently.

Voice features are also coming to Gmail, Docs, and Keep, allowing business users to brainstorm, organize, and complete tasks hands-free inside Workspace apps.

For developers, Google has announced Managed Agents API on Agent Platform. The API allows teams to create custom agents that reason, call tools, and execute code inside Google-hosted remote environments through a single API call.

Google is also integrating CodeMender into Agent Platform. The AI code security agent, originally developed by Google DeepMind, is designed to identify vulnerabilities, recommend fixes, test changes, and apply patches with developer approval.

Several Gemini Enterprise customers are already testing CodeMender, with wider availability expected later. Google is also rolling out an AI Content Detection API on Agent Platform to help businesses identify AI-generated content from Google and other popular models.

The next release points are already set. Gemini Omni Flash is due in the coming weeks, Gemini 3.5 Pro is expected next month, Workspace AI previews are planned for the summer, and Antigravity is scheduled to reach Gemini Enterprise in the coming months.