Google has released Gemini 4 Argon, which it describes as its most capable model for long-horizon work such as software engineering, enterprise knowledge tasks and cybersecurity defence. The model card was published on October 13, 2026, and access is rolling out across Google’s products at once.
Who can use it, and where
- Developers: through the paid tier of the Gemini API in AI Studio and Android Studio.
- Google AI Ultra subscribers: in Google Antigravity, Google’s agent-first development platform, and in AI Studio.
- Everyday users: in the Gemini app and in AI Mode in Search, for Google AI Ultra and Pro subscribers.
Google also says Antigravity’s agentic coding interface is now available directly inside the Gemini macOS app, which narrows the gap between chatting and writing code. Gemini 4 Argon is not available on the free tier.
The specification that matters most
The headline change is a unified context window of 1,048,576 tokens that input, thinking and output all share, plus long decode that allows a single response to run past 64,000 tokens and up to the remainder of that window. Google also exposes tunable thinking levels — low, medium and high — with medium as the default.
Practically, a unified window means a very long input leaves less room for a long answer, and vice versa. That is a real constraint for anyone feeding in large repositories or document sets and expecting a lengthy response, and it is the kind of detail that determines whether a model is usable in a specific workflow.
Pricing and the deadline that matters
Argon is available at introductory pricing through November 30, 2026: $2.00 per million input tokens and $10.00 per million output tokens. From December 1, 2026 those rates double to $4.00 and $20.00. Implicit caching discounts cached input by 95%, to $0.10 per million tokens during the introductory period and $0.20 after. Anyone building a product on top of the model should model both price points rather than the launch rate alone.
What Google claims, and what it does not
Google’s product posts emphasise behavioural improvements — answering more directly with less hedging, repeating itself less, avoiding clichés, and reporting progress more often on longer tasks — alongside gains in coding and agentic work. Agentic video understanding is now generally available, and the Antigravity agent in Gemini Managed Agents supports Argon.
Independent benchmarking against competing frontier models was not part of this article’s research, so treat the capability claims as vendor statements. The verifiable facts are availability, context limits, thinking levels, pricing and dates, all of which are documented in Google’s developer material.
Sources
- Google blog, “Gemini 4 Argon is now available for everyone”: https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-argon/
- Google blog, “Gemini 4 Argon is now available in the Gemini app”: https://blog.google/innovation-and-ai/products/gemini-app/gemini-4-argon-availability/
- Google blog, “Gemini 4 Argon is now available for developers”: https://blog.google/innovation-and-ai/technology/developers-tools/build-with-gemini-4-argon/
- Google AI for Developers, “What’s new in Gemini 4 Argon”: https://ai.google.dev/gemini-api/docs/latest-model
- Google DeepMind, Gemini 4 Argon model card (published October 13, 2026): https://deepmind.google/models/model-cards/gemini-4-argon/
