Google introduces Gemini 4 Neon model for coding and agentic tasks
Google has released Gemini 4 Neon, a new model in the Gemini 4 family designed for programming and tasks that require working with tools and executing chains of actions, the company announced on its blog.
In the lineup, Neon sits between Carbon, which is designed for complex long-term projects, and the fast Hydrogen for everyday queries. According to Google, at the maximum reasoning level xhigh, the model scored 61 points on the Artificial Analysis Intelligence Index, taking third place, while it took first place on the Artificial Analysis Coding Agent Index with 58.6 points. It scored 65.2% on Terminal-Bench 4.1, took second place on the Vals Index with 70.9%, and placed first in the Vending-Bench 2 business management simulation with a total of $16,843.
The model's context window is 1 million tokens, and the reasoning depth can be selected from four levels, from low to xhigh. Via the API, 1 million input tokens cost $3, output tokens cost $15, and cached input tokens cost $0.15. Google claims that at the standard level, a task on Neon is 32% cheaper than on Carbon, and 50–80% cheaper compared to leading competitors' models in third-party benchmarks.
Neon is available to developers in the Gemini API, Google AI Studio, and Android Studio, and to enterprises in Gemini for business. Subscribers to the AI Pro and AI Ultra plans can select the model in the Gemini app and in Google Antigravity.
Separately, the company announced Neon for cyber, a version designed for finding and fixing vulnerabilities in code. According to Google, it achieved 90.5% on SEC-bench Pro and 76% on AutoBaxBench. Verified security specialists from the Fairwind program will gain access to it in the coming days.
What it means
The model is already available via the Gemini API, Google AI Studio, and Android Studio; the price is $3 per 1 million input tokens and $15 per 1 million output tokens. In the Gemini app, it is available to AI Pro and AI Ultra subscribers.
Sources
- Хабр: Google представила Gemini 4 Neon — модель для кодинга и агентных задач media outlet
- Kod.ru: Google выпустила Gemini 4 Neon для сложного кода и ИИ-агентов: что нового media outlet
- Hi-Tech Mail: Google выпустила Gemini 4 Neon для программирования и анализа данных media outlet
- 3DNews: Google выпустила Gemini 4 Neon — почти флагманский рассуждающий ИИ стал быстрее и дешевле media outlet
This article was prepared with the help of AI and translated from the Russian original. Spotted an error — let us know.
Published 19 November 2026, 15:46 · No updates · Russian original




Comments on this story live under its post in our Telegram channel. Comment on Telegram