Three Times Faster and 60% Lower Costs: Google Unveils AI Model Gemini 4 Neon
Google has introduced Gemini 4 Neon for programming, analytics, and multi-step tasks. According to the company, it is three times faster than Gemini 4 Carbon, with task execution costs 60% lower.

Gemini 4 Neon is already available to subscribers of Google AI Pro and Ultra in the Gemini app. Developers can access it via a paid API, and the model is also integrated into Google Antigravity and the company’s corporate services.
However, individual Neon tokens are more expensive than those of Carbon under current promotional pricing. A million input tokens via the API cost $3, while output tokens cost $15. For Carbon, these prices are $2 and $10, respectively; the discount is valid until November 30.
In the Artificial Analysis Intelligence Index test, the model scored 61 points, and in CursorBench 4.1, it achieved 65.4%. Google also presents results from internal trials, where Neon outperformed Carbon in certain metrics. An independent assessment of the model’s capabilities is not yet available.



