Google designed Gemini 4 Neon to pair the high-end reasoning of its top-tier Carbon model with greater speed and lower cost. [7, 13] The model targets complex, high-frequency tasks in software engineering, scientific research, and knowledge work. It can process text, images, audio, and video. [6, 8, 9]
Google is rolling out Gemini 4 Neon to developers through the Gemini API and offering it in the Gemini app to paid subscribers. [6] Neon costs less than Google's Carbon model. The launch extends Google's rapid pace of model releases as it competes with Microsoft and OpenAI. [8, 13]