Google has released Gemini 4 Neon, a new AI model built for programming, analytics, and multistep tasks. According to the company, it matches the output quality of the heavier Gemini 4 Carbon model while running three times faster. The system is now available to paying subscribers and enterprise developers through the Gemini app and a paid API.
Google designed Gemini 4 Neon to process text, images, audio, and video across a 1,000,000 token context window. In internal benchmarks, the model scored 61 points on the Artificial Analysis Intelligence Index and 65.4 % on CursorBench 4.1. The combination of speed and memory allows developers to work with large codebases without long processing delays.
Google says Neon reduces overall task cost by 40 % compared with Carbon, but its raw API token rates are higher at the moment: $3 for 1 million input tokens and $15 for 1 million output tokens, versus Carbon’s $2 and $10, respectively, which reflect a temporary promotional discount.
Neon is available through multiple channels. Google AI Pro and Ultra subscribers can select Neon in the Gemini interface. Developers can access it via the paid API, and enterprise customers will find it integrated into Google Antigravity and Google Cloud services.

