Google Gemini 4 Argon Processes Nearly 25 Trillion Tokens in First 72 Hours

Google's Gemini 4 Argon model processed nearly 25 trillion tokens during its first 72 hours of general availability, roughly tripling the volume Gemini 3 handled over the same timeframe. Following an initial September 30, 2026 release to trusted cybersecurity partners and a wider October 13 launch, Argon expands the input-output limit to 1M tokens compared to Gemini 3's 64K limit. Internal benchmarks reportedly showed the model increased coding task submissions by over 35% while reducing code rollbacks across long-horizon reasoning workflows. API access currently costs an introductory $2 per million input tokens and $10 per million output tokens, with both rates scheduled to double after November 30, 2026. Google CEO Sundar Pichai acknowledged ongoing computing capacity constraints, and the company expects demand to outrun supply as enterprise adoption grows.