Google Gemini 4 Neon: New AI Model Promises Three Times Faster Performance

In Short

Google’s new Gemini 4 Neon is aimed at developers and businesses, offering faster performance, advanced reasoning capabilities, competitive benchmark results and lower costs.

Google Gemini 4 Neon
X

Google Gemini 4 Neon

Font size
FOLLOW ON Google News

Google has added a model to its Gemini 4 AI lineup. The new model is called Gemini 4 Neon. Gemini 4 Neon is made to handle hard software workflows. I notice that Gemini 4 Neon can do this without the costs that come with Google’s flagship product.

Google unveiled Gemini 4 Neon on November 18 2026, just after Google introduced Gemini 4 Carbon and launched Gemini 4 Hydrogen on November 12. Gemini 4 Neon sits between Gemini 4 Carbon and Gemini 4 Hydrogen. Gemini 4 Neon is intended to give reasoning capabilities for tasks, like programming, research, data analysis and other complex operations.

Google says this new model operates three times faster than Gemini 4 Carbon, while reducing overall costs by about 60%. With its arrival, the company has replaced Gemini 4 Argon within the Gemini application and retired the previous version, Gemini 3.1 Pro.

What sets Gemini 4 Neon apart?

Google’s latest model fills the gap between its lightweight AI offerings and its flagship product. While Gemini 4 Hydrogen focuses on quick, everyday queries, Carbon is designed for more demanding workloads and large-scale projects.

Neon combine the strengths of both: it offers advanced reasoning for professional tasks while prioritizing faster response times and lower operating costs.

The model supports text, audio, image, and video inputs. Furthermore, developers can choose from four levels of reasoning—low, medium, high, and extra-high—depending on the complexity of the task and the required response speed.

Performance test results place Neon among the top models

Neon’s performance has also gained attention in independent evaluations. The platform Artificial Analysis awarded it 61 points on its "Intelligence Index" using the "extra-high" reasoning level, placing it third overall. It ranked behind Anthropic’s Claude Fable 5.2, which scored 65 points, and OpenAI’s GPT-6.2 Astra, with 62 points. Neon managed to outperform Gemini 4 Carbon, which recorded 60 points. The model also delivered notable results in science and mathematics, achieving 70.5% on Terminal-Bench Science, 84% on RiemannBench, and 71.3% on HLE-Diamond.

Focus on enterprise and cybersecurity

On security it scored 90.5% on SEC-bench Pro for spotting software bugs and 70.7% on CWE-bench v2, for fixing vulnerabilities.

Google security assessments gave a score of 99.2 percent and Google had a 0.09 percent rate of misaligned actions.

Pricing and availability

Gemini 4 Neon is priced at $3 per million input tokens, $0.15 per cached input token, and $15 per million output tokens.

There is a time-based difference in pricing that's important to mention. Gemini 4 Carbon has a price of $2 for input tokens and $10, for output tokens until November 30. On December 1 these prices will go up to $4 and $20.

Users who have a Google AI Pro subscription or a Google AI Ultra subscription can use Neon through the Gemini app. They need to go to the app and choose the "Deeper analysis" option.

Developers can also test the model using Google AI Studio, Android Studio, Gemini for Business and Google Antigravity. A version that focuses on cybersecurity is expected to be available, for members of the Fairwind program.

Kahekashan is a passionate technophile with a keen eye for cutting-edge gadgets, emerging technologies, and everything in the digital realm. Raised in a Defence family with strong values and a background in literature, she has consistently pursued excellence in every endeavour. Her last full-time assignment involved content writing with the Indian School of Business.

Next Story
Share it