Google's new Gemini 4 Argon model processed nearly 25 trillion tokens during its first 72 hours of general availability, which began on October 13, 2026. This volume is three times higher than the traffic recorded for the previous Gemini 3 model during its opening weekend.
The model features a 1 million token context window, a significant increase from the 64K limit in Gemini 3. Internal data from Google indicates that the model improved coding task submissions by more than 35% and reduced the number of required rollbacks.
Developers currently access the model via API at an introductory rate of $2 per million input tokens and $10 per million output tokens. These prices are scheduled to double after November 30, 2026.
CEO Sundar Pichai noted that Google faces potential computing capacity constraints due to high demand for the new model. The company initially launched Argon for cybersecurity partners on September 30, 2026, before the wider release.
