Google Launches Gemini 4 Argon with 1M Token Capacity for Complex Engineering Workflows

Google DeepMind unveiled Gemini 4 Argon, a frontier AI model engineered for specialized coding, deep research, and complex engineering tasks, with an industry-leading 1 million token output capacity. The model marks a significant expansion from Gemini 3's 64K token limit, enabling professionals to tackle extended problem-solving workflows in software engineering, quantum computing optimization, and enterprise knowledge work.