Google Launches Gemini 4 Argon, Its Most Advanced AI Model Yet, With 1M Token Output Limit

Google has released Gemini 4 Argon, a frontier AI model designed for complex long-horizon workflows including software engineering, cybersecurity defense, and enterprise knowledge work, now appearing on Google Cloud Platform quotas pages signaling imminent wider API availability. Announced on September 30, 2026, Gemini 4 Argon represents a substantial leap in capability, featuring an industry-leading 1 million token output limit—a 16-fold increase from the previous 64,000 token cap—alongside expanded reasoning and coding performance.