Google Unveils Gemini 4 Argon for Robotics and Global AI Access
Google’s new AI model targets robotics control and advanced coding, offering discounted API rates and app access to subscribers.
Key points
- Gemini 4 Argon is now available globally via the Gemini app, Search, and developer APIs.
- Launch pricing offers temporary discounts on input and output tokens compared to standard rates.
- The model supports robotics simulations, coding tasks, and advanced reasoning capabilities.
Google has made its Gemini 4 Argon artificial intelligence model available to paying users around the world. The system operates within the Gemini application, through Search results, and via developer application programming interfaces. This release allows consumers and businesses to access advanced reasoning and coding features without waiting for further regional rollouts or exclusive beta periods.
In plain English
Artificial intelligence models like Gemini 4 Argon process information to perform specific tasks. These models can write code, analyze complex documents, or control physical robots. Google designed this specific version to handle heavy computational workloads efficiently. It aims to reduce the time developers spend on repetitive coding tasks while improving the accuracy of automated systems.
The background
Google DeepMind product lead Tulsee Doshi described the model as a powerhouse during a discussion with Business Insider. She noted it ranks first on the Vals Index, which evaluates performance in finance, coding, legal, and tax domains. The company claims the model helps verify scientific conjectures and teaches humanoid robots new skills by observing video demonstrations. Internal testing showed reduced errors in coding shortcuts and improved safety metrics.
What changes now
Developers can now access the model at a discounted rate of two dollars per million input tokens. Output tokens cost ten dollars per million during this initial period. These prices are lower than the standard rates of four dollars and twenty dollars respectively. The discount lasts for at least one month. Consumers need a Pro subscription starting at nineteen dollars ninety-nine cents monthly to use the app features.
Key terms explained
Artificial intelligence models process vast amounts of data to generate text, code, or images. Tokens are small units of text that developers use to measure usage costs. Application programming interfaces allow other software to communicate with these models. Robotics simulations create virtual environments where machines can practice tasks before operating in the real world. Teams usually monitor these systems to ensure they follow safety guidelines and produce accurate outputs.
What to do and how to stay safe: Google
- Review your organization’s budget for artificial intelligence usage to account for potential token costs.
- Evaluate whether current security policies cover external API integrations for sensitive data processing.
- Test new models in isolated environments before deploying them to production systems.
- Monitor vendor communications for changes in pricing or feature availability during launch periods.
Step-by-step guide: Stop AI Chatbots Leaking Sensitive Business Data
General security guidance from the Vector Update newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
How much does Gemini 4 Argon cost for developers?
Developers pay two dollars per million input tokens and ten dollars per million output tokens during the launch discount period.
Is the discounted pricing permanent?
No, the launch discount runs for at least one month, after which standard rates of four dollars and twenty dollars apply.
What subscription level is needed for consumers?
Consumers need a Pro subscription, which starts at nineteen dollars ninety-nine cents per month, to access the model in the app.
