Google's Gemini 4 Argon is live — intro pricing doubles November 30

Google's developers blog posted October 13, 2026, at 11:00 AM CT, makes it official: starting today, developers can build with Gemini 4 Argon through the paid tier of the Gemini API in Google AI Studio and Android Studio. Google AI Ultra subscribers can also try it in AI Studio and in Google Antigravity, the company's agent-first development platform.

For subscribers, Argon rolls into the Gemini app and AI Mode in Search for AI Pro ($20/month) and AI Ultra (from $99/month) plans; inside Antigravity it stays Ultra-only, per Reuters and Android Headlines. Intro API pricing is $2 per million input tokens and $10 per million output tokens, with cached input at 95% off the input rate — but that promotional rate runs only through November 30, after which the standard price is $4 and $20.

Argon is not new — Google announced it on September 30, in a post from Koray Kavukcuoglu, SVP of Google DeepMind and the company's chief AI architect. But the September release went only to trusted cyber defenders in Google's Fairwind Program. Google said safely releasing frontier capabilities at this level required a phased approach, and that it was engaged in the U.S. government's voluntary pre-release model access process.

Google is now opening the Fairwind Program to hundreds more vetted cybersecurity defenders — giving them a version of Argon without its cyber guardrails — while offering 10,000 scientists six months of free access to Antigravity's highest usage tier. The model is built for long-horizon workflows needing deep reasoning and structural precision: code migrations, frontier research, legal and finance work, and cybersecurity defense. It takes text, images, video, audio, and PDFs as input but produces text only, with a unified 1,048,576-token context window shared by input and output.

Google's internal-use claims are the interesting part, and they are Google's claims, not independent results. Teams using Argon in Antigravity increased agentic code submissions by more than 35% while cutting code rollbacks by nearly a third, Reuters reports. Google engineers used it to clean up six million lines of legacy code in the Chrome browser in weeks instead of months, eliminating an entire category of browser crashes — and over at YouTube, it spotted redundant emoji rendering code, halving the frontend comment server load and cutting client latency.

On paper, Google says Argon ranks first on the Vals Index for legal, financial, and coding tasks, hit 77.9% on DeepSWE (against 74.2% for Claude Opus 5.5 and 74.1% for GPT-6 Astra), and scored 87% on an Embodied Reasoning benchmark — including watching a video of a person playing cornhole and driving an Apptronik Apollo humanoid to sink seven consecutive shots. Early enterprise testers include Deloitte, LexisNexis, Ambience Healthcare, and Isomorphic Labs. None of the benchmark claims have been independently verified.

If you want to try it today, the routes are: the paid tier of the Gemini API in Google AI Studio or Android Studio for developers, or an AI Pro or Ultra subscription for the Gemini app and AI Mode in Search. Start from Google's Gemini API Developer Guide and the Antigravity blog for setup. And if you are budgeting API spend, get your experiments in before November 30 — the token price doubles after that.

One honest caveat before you commit budget: the benchmark wins and the internal success stories all come from Google. The price window, the access tiers, and the Fairwind expansion are facts; everything else is a claim to verify against your own workloads.