Short answer: On 8 October 2026, at Gemini at Work 2026, Google Cloud announced the Gemini agent, a single universal agent for work. You give it objectives rather than step-by-step instructions; it plans the work, runs it in the cloud across Gemini and Claude models, and works inline in Workspace, Slack and Microsoft 365. It is in private preview.
By Agent Playbox Editorial · Published
Gemini Agent: Google Cloud's universal agent for work
Key facts
- Announced 8 Oct 2026 at Gemini at Work 2026 as a universal agent for work (blog.google)
- 'You give it objectives, not instructions' (Thomas Kurian keynote, adapted by Google Cloud) (chromeosphere.com)
- Runs each job on the model that fits best: Gemini and Anthropic's Claude models today (9to5google.com)
- Coworker agents get their own Workspace account: email, calendar, Drive and a directory entry (chromeosphere.com)
- Works inline in Gmail, Docs, Sheets, Slides and Chat via @Gemini (9to5google.com)
- Spend caps in the Cloud Billing Console pause a project's agent when reached (chromeosphere.com)
- Private preview; wide availability soon for select Workspace Business and Enterprise plans (9to5google.com)
What did Google Cloud announce on 8 October 2026?
At Gemini at Work 2026, Google Cloud introduced the Gemini agent, “a universal agent for work” (Google blog). From one prompt box it handles knowledge work, answers questions, creates media and writes and runs code, then brings back something finished inside the documents, inbox and developer tools people already use.
Google Cloud CEO Thomas Kurian summed it up: “You give it objectives, not instructions. You delegate an outcome and come back to finished work” (keynote adaptation). It can be your personal assistant or a team member, such as a project manager or a finance analyst.
How does the Gemini agent work?
Persistent work in the cloud
The agent runs in the cloud with one set of memories and context across every device, so work that takes hours or days keeps going after you close your laptop. It can spin up sub-agents for multi-step jobs, as 9to5Google reports, and keeps four kinds of memory: session, semantic, procedural (including skills it writes for itself) and episodic.
Model choice
“Gemini is the agent, and the model underneath it is a separate choice.” It runs each job on the best-fitting model across Google’s Gemini family and Anthropic’s Claude models today, with other models planned. Smart Routing sends simple tasks to lighter models and hard ones to stronger models.
Everywhere you work
Google lists web, iOS, Android, Windows and Mac, plus the command line, Google Workspace, Microsoft 365 and Slack. In Gmail, Docs, Sheets, Slides and Chat you mention @Gemini. Via tools and MCP it reaches systems such as Salesforce, Jira and BigQuery, and an API lets it run headless inside other apps.
What are Gemini coworker agents?
You describe a role, say an events coordinator, and Gemini creates a coworker agent with its own Workspace account: email address, calendar, Drive and a place in the company directory (Implicator reports @agents.company.com addresses). Colleagues add it to a Chat space or @mention it in a document comment, and its edits appear under its own name in version history. It acts under its own identity and sees only what the team shares with it.
How do security and cost controls work?
Every agent gets its own cryptographically attested identity with least-privilege permissions, and every action lands in an audit trail attributed to the agent. Agents run in an Agent Sandbox, and all traffic passes through Agent Gateway, which applies company policies across every agent at once.
For budgets, teams set a hard spend cap per project in the Cloud Billing Console. When it is reached, that project’s agent pauses and resumes with one click, and costs can be charged back to departments.
When is it available and what does it cost?
The Gemini agent is in private preview, with wide availability “soon” for select Workspace Business and Enterprise plans, per 9to5Google. Implicator reports that Google says it will come at no additional charge for Gemini Enterprise customers, including existing Workspace business and enterprise subscriptions, and that executives told analysts to expect general availability around the end of October or early November. Whether persistent coworkers or long autonomous tasks cost extra is still unknown; see the Gemini Agent page.
What does the Gemini agent mean for teams?
Workspace teams get an agent inside the tools they already use. Google’s examples: when a manager emails asking for a project update as a slide deck, Workspace offers a one-click option to delegate it; ask Gemini to set up a meeting with “the usual team of regional event leads” and it works out who they are, checks calendars and starts the thread.
Practical starting points include a weekly investor update draft, competitor watch and lead research before calls; our meeting notes and follow-ups template makes a good first coworker role.
How does the Gemini agent compare with OpenAI Dots and Grok Bot?
It arrives nine days after OpenAI launched Dots, always-on agents with their own cloud computer, and alongside Grok Bot and Claude Cowork. Gemini’s strengths are deep Workspace integration, multi-model routing and coworker agents with full Workspace identities. Compare them in Gemini Agent vs OpenAI Dots, Gemini Agent vs Grok Bot and business vs personal agents.
Timeline: how did we get here?
- : Google Cloud launches the Gemini agent, a universal agent for work. At Gemini at Work 2026, Google Cloud announced the Gemini agent: a single agent for work that takes objectives rather than instructions, runs jobs across Gemini and Claude models and works inline in Workspace, Slack and Microsoft 365.
- : OpenAI launches Dots, always-on agents with their own computer. At DevDay, OpenAI introduced Dots: always-on agents powered by GPT-6 Astra, each with a cloud computer and access to 4,000+ apps via plugins.
Agent snapshot: what do we know so far?
Facts below come from the agent profiles in our directory, each linked to its source.
Gemini Agent at a glance
- Vendor: Google Cloud
- Where you use it: Web, iOS, Android, Windows and Mac desktop, command line, Google Workspace, Slack, Microsoft 365 (per Google's keynote adaptation and 9to5Google)
- Pricing and availability: Private preview. Included at no additional charge for Gemini Enterprise customers, including existing Workspace business and enterprise plans (per Implicator); charges for persistent coworkers or long autonomous tasks not yet published
- Funding: n/a (Google Cloud product)
OpenAI Dots at a glance
- Vendor: OpenAI
- Where you use it: ChatGPT (web, desktop, mobile), Codex, Slack, Microsoft Teams
- Pricing and availability: Included in ChatGPT Pro and Business Premium (eligible markets); Enterprise beta. Pro access excludes the EEA, UK and Switzerland at launch (per The Decoder)
- Funding: Unknown
Related comparisons and guides
Agents in this story
Google Cloud's universal work agent: give it objectives, not instructions. It works across Workspace and third-party channels and can create coworker agents with their own Workspace identity.
Always-on agents powered by GPT-6 Astra, each with its own cloud computer, working towards your goals 24/7.
Frequently asked questions
What is the Gemini agent?
A single, universal agent for work from Google Cloud, announced on 8 October 2026. You give it objectives and it plans and completes the work across your tools.
Which AI models does the Gemini agent use?
It picks the best-fitting model per job, across Google's Gemini models and Anthropic's Claude models today.
What is a Gemini coworker agent?
An agent with a defined role and its own Workspace account (email, calendar, Drive, directory entry). It acts under its own identity.
How much does the Gemini agent cost?
Google says it is included at no additional charge for Gemini Enterprise customers (per Implicator). Charges for persistent coworkers or long tasks have not been published.
Sources
Updated