Google releases Gemini 4 Argon for users and developers

Google began rolling out Gemini 4 Argon globally on October 13, 2026. Developers can access the model through the paid Gemini API in AI Studio and Android Studio. Google AI Pro and Ultra subscribers gain access in the Gemini app and AI Mode in Search, while Ultra subscribers can also use it in Google Antigravity. Google is extending access to businesses and researchers as well.

The launch follows Google's September 30 announcement, when access was limited to trusted cybersecurity defenders. Today's change makes Argon available for people to test on their own work. Google positions it for demanding programming, document analysis, finance, legal work and science. Through the API, the model has a shared input and output limit of one million tokens, giving extended tasks more room.

Introductory API pricing is $2 per million input tokens and $10 per million output tokens, with cached input discounted by 95 percent. Google says this introductory period ends on November 30, 2026; subsequent prices will be $4 and $20 respectively. These are usage rates, so the cost of completing a task also depends on how many tokens it consumes.

Google illustrates Argon's capabilities with internal engineering projects. The company reports faster code refactoring in Chromium and optimizations to its Python Agent Development Kit. It also describes experiments in quantum computing and enzyme design. These examples support Google's case for longer, multi-step work, but remain provider-reported results rather than a guarantee of comparable gains elsewhere.

The release also expands Google's Fairwind Program for cybersecurity partners. Separately, 10,000 scientists can apply for six months of free access to Antigravity's highest usage tier, subject to eligibility, selection and capacity. Google describes safeguards against harmful requests, malicious instructions hidden in external content and agents taking unauthorized shortcuts.

For users and makers, the practical significance is access: Argon can now be compared on real tasks alongside other models. Businesses will need to measure completion quality, total cost and the amount of human review required. A large token allowance can accommodate extensive work, but does not establish that every step is accurate. The launch turns an earlier capability announcement into a product that more people can evaluate.