Argon was first announced on September 30 to a small group of trusted cyber defenders in Google's Fairwind program. As of October 13, developers on the paid Gemini API tier can call it, Ultra subscribers get it in AI Studio and Antigravity, and Pro and Ultra subscribers can use it in the Gemini app and AI Mode in Search. Google reports 77.9% on DeepSWE v1.1 and raised the output limit to 1M tokens from 64K. Reuters lists Deloitte, LexisNexis, Ambience Healthcare and Isomorphic Labs among early testers. The release follows a rough stretch: Google said in September it would not ship Gemini 3.5 Pro, which had been promised for June, and Reuters reports DeepMind had a leadership shakeup. For builders, the practical points are the price clock and the long output window. The introductory rate reportedly rises after November 30, so cost models need a second scenario. A 1M-token output ceiling changes what one agent run can attempt, but benchmark claims are Google's own and worth testing on your workload.