Google launches gemini 4 argon with 1 million token output limit

Google has launched Gemini 4 Argon, its most advanced AI model, offering developers and enterprises access through AI Studio and API platforms. The model supports Google AI Ultra subscribers and is available to users of Gemini applications and search AI modes. It features a 1,048,576 token (approximately 1 million) output limit, with input, thinking, and output sharing the same context window. Pricing starts at $2 per million tokens, with exchange rates set at 6.6921 RMB per USD as of October 14, 2026. Gemini 4 Argon excels in complex tasks across coding, finance, law, cybersecurity, and science, outperforming other models in benchmark tests. It ranks first in the Vals Index, which measures performance in critical economic areas. The model is already being used by internal teams and external partners to optimize workflows, identify security vulnerabilities, and advance research in quantum computing, nuclear fusion, and life sciences. Google has provided 10,000 scientists with free access to the highest usage tier for six months. The company has also expanded the Fairwind project to help cybersecurity professionals. Enhanced security measures, including improved resistance to jailbreaking and prompt injection, have reduced harmful request rates by nearly half.