Google presented on Wednesday (18) Gemini 4 Neon, a new artificial intelligence model developed for complex programming, data analysis, scientific research and enterprise operations tasks. The proposal is to combine advanced reasoning capability with greater execution speed and cost efficiency.

Within the Gemini 4 family, Neon occupies an intermediate position between Carbon, aimed at larger and longer projects, and Hydrogen, focused on the fast execution of everyday activities.

In independent evaluations by Artificial Analysis, the model reached 61 points on the Intelligence Index, taking 3rd position overall at the time of launch, behind GPT-6.2 Astra and Claude Fable 5.2. The result was obtained in Neon's highest reasoning configuration, called xhigh.

The model offers four reasoning levels, allowing computational effort to be adjusted according to the complexity of the task. It also features a 1 million token context window, shared between input and output, with support for processing text, images, audio and video.

Gemini 4 Neon pricing and availability

On the API, Gemini 4 Neon costs US$ 3 per million input tokens and US$ 15 per million output tokens. Processing input tokens stored in cache costs US$ 0.15 per million, a 95% reduction compared with the standard input rate.

Availability began with Google AI Studio, Android Studio and Gemini for business. Users on the Google AI Pro and Ultra plans can also access the model through the Gemini app and the Google Antigravity development platform.

Tools mentioned

More from Radar