Google puts Gemini 4 Argon in its app, adds 40-plus partners

Key Points

1. Google said Gemini 4 Argon is now live in the Gemini app for paying subscribers.
2. The model answers more directly and shares progress during longer tasks, according to Google.
3. The release pushes Gemini from chat assistant toward an agent executing work across outside services.

The latest:

Gemini 4 Argon is now available inside the Gemini app, Google announced on Oct 13, 2026, describing it as a frontier reasoning model with stronger performance on complex tasks. The company said the model hedges less, repeats itself less and drops clichés from its answers. Access is limited to Google AI Pro and Ultra subscribers, starting the same day.

Details:

• The model: Google said Argon delivers frontier reasoning and improved performance on complex tasks, answering questions more directly with less hedging. The company also said it avoids unnecessary repetition and skips clichés across responses. Google did not publish benchmark figures or name the models it was compared against.
• Long tasks: For longer jobs, Google said Argon keeps users in the loop by sharing its progress more often and more effectively. The company framed that as a behavioral change in how the model reports back during extended work, rather than a separate feature users switch on.
• Coding: Google said Argon brings major improvements in coding and that the company is making it easier to move between chatting and coding. Google Antigravity’s agentic coding interface is now available directly inside the Gemini macOS app. Google did not say whether the same interface is coming to Windows or other platforms.
• Chat to tasks: A toggle in the main chat window now switches Gemini from chat to tasks. Google said that when users choose tasks, Gemini can work on their behalf even when they are away from their computer, including across Google apps.
• The integrations: Gemini now connects to more than 40 apps, according to Google, with Instacart, OpenTable, Adobe and Shopify named as examples. The company cited ordering groceries, booking tables, designing and building a business as sample uses. The full list of the 40-plus partners was not published in the announcement.
• Who gets it: Google said Google AI Pro and Ultra subscribers can now access Gemini 4 Argon and assign Gemini tasks, with the updates available starting today. The announcement set no date for wider availability to free users.
• Rollout framing: The announcement was filed under the Gemini App and Google One categories, placing the release inside Google’s paid subscription stack rather than presenting it as a standalone model launch. No pricing change was announced alongside it.
• The context: The release follows a separate Google announcement of a new wave of Connected Apps rolling out to Gemini, part of the same push to link the assistant to third-party services. Google has also been promoting Gemini for everyday consumer uses, including household tasks.

Between the lines:

The headline change may be less the model than the surface around it. A toggle from chat to tasks, more than 40 connected services and an agentic coding interface inside the desktop app all point the same way: Google is positioning Gemini to execute actions in outside products, not just answer prompts. Gating both Argon and tasks behind Pro and Ultra ties that capability directly to subscription revenue.

What’s next

Watch whether Google publishes benchmark results for Argon, releases the full list of more than 40 partner apps, extends the Antigravity coding interface beyond macOS, or opens Argon and tasks to free-tier users.