Google released Gemini 3.7 Flash on August 13, its newest AI model aimed at software coding, web development, and multi-step business automation. The launch comes just three weeks after Gemini 3.6 Flash, continuing Google's accelerated release pace even as its flagship Gemini 3.5 Pro model remains delayed.
The new model posts sharp benchmark gains over its predecessor: the DeepSWE coding score jumped from 49.0% to 65.3%, FrontierCode rose from 34.4% to 43.6%, and its WebDev Arena Elo climbed from 1538 to 1588. Google also reports stronger multi-step planning, debugging, and instruction-following for autonomous agent workflows.
Pricing is the headline change: through the end of 2026, Gemini 3.7 Flash costs $0.75 per million input tokens and $3.75 per million output tokens, half the launch price of its predecessor. It is rolling out now in Gemini Spark, Google Antigravity, AI Studio, Android Studio, and the Gemini Enterprise Agent Platform.
