Understanding Google’s Gemini Model Names

Google named its Gemini models after chemical elements. Here’s what Hydrogen, Neon, and Carbon each do.

Learn Prompting Newsletter

Your Weekly Guide to Generative AI Development

GenAI News

Understanding Google’s Gemini Model Names

Google named its Gemini models after chemical elements. Here’s what Hydrogen, Neon, and Carbon each do.

Hey there, 

Since mid-October, Google has released four Gemini 4 models: Argon, Carbon, Hydrogen, and now Neon, which launched on Nov 18. That’s a lot of new names in a short stretch, and it’s easy to lose track of which one does what and which one you actually have access to. So this week, we’ll sort out how the four fit together and take a closer look at what Neon can do.

How the Four Gemini 4 Models Fit Together

Google has moved away from its old Flash and Pro names in favor of a new system. All four are named after chemical elements, and the names don’t hint at which is which. Here’s the lineup as of Nov 20.

Hydrogen. Google’s fast, everyday model, released Nov 12. It’s in the Gemini app for everyone, though people on the unpaid plan get it through automatic model selection. Think of it as the new “Flash” model, built to be fast, efficient, and inexpensive.

Neon. Released Nov 18, and Google says it balances depth and efficiency for complex work like coding, analysis, and science. It sits between Hydrogen and Carbon, and for Pro subscribers, it’s taken the old Pro model’s place in the Gemini app. It’s available to Google AI Pro and Ultra subscribers.

Carbon. Google’s model for the hardest, long-running projects. It’s available to Google AI Ultra subscribers.

Argon. Google’s frontier model from mid-October. For Pro subscribers, Neon has now replaced it in the Gemini app’s model picker.

In other words, think of Hydrogen as the everyday option, Neon as the step up, and Carbon as the top tier.

What Neon Can Do

Google describes Gemini 4 Neon as bringing frontier-level performance to complex coding, knowledge work, and science without the wait. Google says it costs 32% less per task than Carbon and has its lowest rate of made-up answers yet, based on its own testing. Tasks in Gemini can now run on it too. As for how it stacks up, Google says it ranks second on an index of more than 4,200 real-world agent sessions, and third on the Artificial Analysis Intelligence Index, just behind larger, slower models.

People have already been putting it to work. Google shared some examples of what the community built with Neon, including a glowing 3D Earth, an interactive black hole with switchable camera views, and a 3D pendulum wave.

An early-access tester had Neon build 17 interactive demos in a day and shared the misses along with the hits. For example, a solar system demo started with outdated moon counts, and Neon updated them after the tester flagged it.

How to Check Which Gemini You Have

Open the Gemini app and tap the model picker to see which models your plan includes. Pro subscribers should see Neon, which the app labels “Deeper analysis,” and Ultra subscribers also get Carbon. The rollout is gradual, so Neon may take a little while to show up on your account.

Build Real Apps with Google AI Studio

Build Real Apps with Google AI Studio is a hands-on vibe coding course that takes you from blank prompt to deployed, shareable app, no coding background needed.

You follow the exact prompts, watch the build happen in real time, and end each lesson with a working app you can share. Each project teaches the judgment calls a live demo skips: when to let the model choose features, what to do when the preview breaks, and how to publish so others can use what you built.

Taught by Chandler Kilpatrick, Head of Education & Content at Learn Prompting, who has taught hundreds of thousands of people how to use generative AI.

My Thoughts

Google is moving fast and continuing to give us new models but there is a catch. I think the pace and the naming are getting in the way for regular users. Argon was Google’s frontier model just over a month ago, and it’s already gone from the Gemini app’s model selector for Pro subscribers. Names like Hydrogen, Neon, and Carbon don’t tell you what each one is for, so you end up having to look them up, which is the opposite of what a model menu should do. It’s telling that Google labels Neon “Deeper analysis” inside the app, since that’s a much clearer description than the name itself. Names like Pro and Flash told users what to expect, while version numbers made it easier to understand which model was the newest. For now though, a cheat sheet like the one above is probably the easiest way to keep up.

Reply

or to participate.