Google released Gemini 4 Argon, its newest frontier AI model, on Oct. 13, making it immediately available worldwide through the Gemini API, AI Studio, Android Studio and the Gemini app, the company said in a blog post.
Argon is built for complex work in coding, finance, legal, cybersecurity and science, according to Google, and supports a 1-million-token combined input and output limit.
The model ranked first on the Vals Index, a benchmark covering finance, coding, legal and tax tasks, Google said, and scored 87% on the company's embodied reasoning benchmark for robotics applications.
Google priced Argon at $2 per million input tokens and $10 per million output tokens through Nov. 30, when the rate doubles to $4 and $20. Developer Simon Willison, who updated his llm-gemini plugin to support the model, said the introductory price matches OpenAI's GPT-6 Sol and undercuts Anthropic's Claude Opus 5.5, though he cautioned that "prices like this aren't really comparable these days" because token efficiency differs across models.
Argon also reaches consumers through the Gemini app and AI Mode in Search for Google AI Ultra and Pro subscribers. For builders comparing frontier model APIs, the pricing window closing Nov. 30 gives a concrete deadline to test the model before the cost of running it rises.