Dynamic Beating AI News Flash: Google CEO Sundar Pichai stated that Gemini 3.7 Flash broke the growth record of the previous Gemini model in its first week online, becoming Google's fastest-growing Gemini model to date. However, Google did not disclose the specific growth metrics and usage scale.
This model has indeed hit the sweet spot in terms of cost-effectiveness. The Artificial Analysis gave it a comprehensive intelligence index of 56, with a high reasoning mode output speed of approximately 340 tokens/s, a cost of around $0.40 per evaluation task, and it has entered the Pareto frontier of intelligence and completion time. The cost of the medium reasoning mode has further dropped to $0.26, entering the Pareto frontier of intelligence and cost.
In independent validation by the ARC Prize, Gemini 3.7 Flash's high reasoning mode scored 84.6% on ARC-AGI-2 (Abstract Reasoning Benchmark) with a task cost of only $0.25.
Google has swiftly integrated 3.7 Flash into its own products. The Gemini App, Antigravity, and AI Studio have already adopted it, and Google Search has also started using this model.

