ROHAN TECH OFFICIAL • Technology • AI • Apps • Tutorialsinfo@rohantechofficial.com
ROHAN TECH OFFICIALTechnology • AI • Apps • Tutorials • Digital World
Home → Tech News → Google Launches Gemini 4 Neon, Its Fast Mid-Tier Model Between Hydrogen and Carbon

Google Launches Gemini 4 Neon, Its Fast Mid-Tier Model Between Hydrogen and Carbon

By Rohan Jaspal • November 19, 2026 • Updated November 19, 2026
Google Launches Gemini 4 Neon, Its Fast Mid-Tier Model Between Hydrogen and Carbon

Google has added a fourth model to the Gemini 4 family in under two months. Gemini 4 Neon, announced on November 18 by Google DeepMind, sits between the heavyweight Carbon and the fast, cheap Hydrogen — and Google pitches it as the one you would actually use all day.

Koray Kavukcuoglu, DeepMind’s senior vice president and chief AI architect, introduced Neon as a model for “frontier-level intelligence with fast end-to-end task completion” across complex software engineering, knowledge work and science. In plain terms, it is meant to handle hard jobs without making you wait like Carbon does.

Where Neon sits in the Gemini 4 lineup

The Gemini 4 family now has a clear pecking order. Carbon takes on the hardest long-horizon projects. Hydrogen handles fast everyday execution. Neon fills the middle: high-frequency complex coding, analytical reasoning and enterprise workflows — the kind of work that needs real brainpower but also needs to finish before lunch.

Neon also takes over Argon’s slot in the Gemini app. Paying subscribers now find it in the model picker under “Deeper analysis”, where Argon used to sit. It is the fourth Gemini 4 model Google has shipped in about seven weeks, which says plenty about the pace DeepMind is working at right now.

For readers keeping track of the family, we covered Gemini 4 Hydrogen when it arrived as the fast option, and Gemini 4 Carbon when it topped the benchmarks. Neon is the model that tries to give you Carbon-grade answers at Hydrogen-like speed.

What the benchmarks actually say

Google points to the independent Artificial Analysis Intelligence Index, where Neon scores 61 at its highest “xhigh” thinking level. That puts it third overall, behind GPT-6.2 Astra and Claude Fable 5.2, and one point ahead of Carbon. On the Coding Agent Index, Neon scores 59 when used through the Antigravity CLI, three points ahead of Carbon.

The speed numbers are where Neon pulls away. Artificial Analysis measured about 110 output tokens per second from Neon, close to double Carbon’s roughly 57. A typical task costs $6.44 with Neon against $7.76 with Carbon — about 17% less — and finishes in around 21 minutes instead of 36, roughly 40% faster. The catch is that Neon uses about 15% more output tokens per task, so the saving comes from cheaper tokens, not fewer of them.

Those are benchmark figures, not a promise about your workload. Still, the direction is clear: Neon trades a small amount of top-end intelligence for a large gain in speed and cost.

Pricing and who can use it

Neon’s API pricing is $3 per million input tokens and $15 per million output tokens. Cached input costs $0.15 per million, a 95% discount. That is 25% below Carbon’s standard $4/$20 rate — and it matters from December 1, when Carbon’s introductory $2/$10 pricing ends and its price doubles. From that date, Neon is cheaper per token than Carbon as well as faster.

There is no free API tier for Neon, and no free tier in the app either. It is rolling out now to Google AI Pro and AI Ultra subscribers in the Gemini app, and to developers through the Gemini API, AI Studio and Antigravity. Like its siblings, it carries a 1 million-token context window, so a whole project — code, tables, PDFs, audio and video — can fit in a single task.

Neon also offers four thinking levels — low, medium, high and xhigh — so teams can trade intelligence against token spend depending on the job. Google is also preparing a cybersecurity version, Gemini 4 Neon for cyber, aimed at finding vulnerabilities and fixing code.

What it means for developers

For most developers, Neon looks like the practical pick of the family. Carbon stays the choice for the hardest single problems, and Hydrogen stays the cheapest way to get quick answers. Neon is the daily driver: repo-scale coding, tool-heavy agents and long-running workflows where waiting 36 minutes per task gets old fast.

The December 1 price change is the date to watch. Once Carbon jumps to $4/$20, Neon becomes the cheaper option per token too, which makes the mid-tier model the sensible default for a lot of teams. If you already pay for AI Pro or Ultra, Neon is sitting in your model picker now — worth a try on your next big coding task before you reach for Carbon.

Google shipped four Gemini 4 models in seven weeks, and also rolled out the Genie 4 world model to select users on the same day. The family is growing fast, and Neon is the one built for everyday heavy lifting.