Gemini 4 Argon debuts with 1M-token output for trusted testers

Google DeepMind introduced Gemini 4 Argon on September 30 as its new frontier model, the term labs use for their most capable systems, and said it was built for complex workflows in coding, enterprise knowledge work and cybersecurity defense. On launch day it was not open to everyone. The announcement says it was rolling out to a set of trusted testers through Google's Fairwind Program, which Google described when it launched the program earlier in September as limited access for governments and trusted partners to its most advanced cyber defense capabilities.

The post itself gives no benchmark results, prices or technical limits. In particular, the one-million-token output figure in this item's headline does not appear in the collected text of the post. The post carried one image, which was not available as text for this explanation, so that figure is unverified here. Broader access came two weeks later, when Google's developer announcement opened the model to developers.