Google launches Gemini 4 Argon with a 1M-token input/output cap
Google has released Gemini 4 Argon, its new flagship model, to developers on the paid tier. The model offers a unified limit of one million tokens across input and output. The launch was announced on Tuesday.
Google’s Gemini 4 Argon arrives as a flagship AI offering for developers on a paid plan. Its one-million-token limit applies across both input and output, meaning the same cap governs what users provide and what the model returns. The Tuesday announcement places the release within the broader AI technology topic, where access tiers and token limits shape how developers can build and test applications.
The release could affect developers and businesses that rely on Google’s AI tools, especially those needing large context windows for complex tasks. Paid-tier access may shape who can experiment with the model, potentially widening gaps between well-funded teams and smaller builders. It may also influence expectations around token limits across the AI industry, though actual societal effects will depend on adoption and use cases.