Have you ever opened a model picker, looked at the options, and wondered why the middle one exists at all? The small one is fast and cheap. The big one is smart and expensive. So what is the one in between for?
On November 18, 2026, Google introduced Gemini 4 Neon, and the middle option is now the interesting one. This is my third Gemini 4 article in nine days, after Carbon and Hydrogen, so let me be clear about what is actually new here. It is not just another benchmark chart. It is where Neon sits in the lineup, what it costs, and one quiet feature change that matters for everyday automation.
A New Middle Tier
Google positions Neon between Hydrogen, which handles fast everyday execution, and Carbon, which tackles the hardest long-horizon projects. Neon is built for complex coding, analytical reasoning, and production-grade enterprise workflows, balancing depth and efficiency.
It has a 1 million-token context window and four thinking levels, so you can dial how much reasoning you pay for. In the Gemini app, Neon replaces Argon in the picker as "Deeper analysis," sitting between "Everyday help" (Hydrogen) and "Code & complex work" (Carbon).
Think of it like coffee sizes. Hydrogen is the small, quick cup. Carbon is the large, strongest one. Neon is the medium that most people actually order, because it is close to the large in strength but easier on the wallet.
Photo by Danial Igdery on Unsplash
The Price Is the Story
Neon costs $3 per million input tokens and $15 per million output tokens, with cached input at $0.15 per million. That is twice the price of Hydrogen, which sounds like a lot until you look at the other end of the ladder.
Carbon and Argon are on promotional pricing at $2 input and $10 output until November 30. After that, they rise to $4 and $20. When the promo ends, Neon at $3/$15 becomes 25% cheaper than Carbon. Google also says Neon costs 32% less per task than Carbon at its default high thinking level. That is Google's own comparison, so treat it as a claim to test, not a promise.
This is the part worth watching. Model pricing is a ladder, and Google just made the middle rung the best value on its own numbers.
What Google Claims It Can Do
Google says Neon ranks #3 overall with a score of 61 on the Artificial Analysis Intelligence Index at its xhigh thinking level, and #1 on the Artificial Analysis Coding Agent Index with a score of 58.6. It also reports 65.2% on Terminal-Bench 4.1.
Here is the honest caveat, and it matters: every one of these figures comes from Google's own post or indexes Google cites. Nobody independent has re-run these tests yet. 9to5Google framed the launch as "Carbon quality at Flash speeds," which is a neat headline, but it is still Google's framing of Google's numbers.
Tasks Can Now Run on Neon
The quieter news is that Tasks in Gemini can now run on Neon. Tasks is Google's feature for handing off multi-step jobs, like analyzing billing statements to spot why a utility charge spiked, or cross-checking receipts and documents to negotiate a refund.
Running those jobs on a faster, cheaper model means they finish sooner and stretch usage limits further, in Google's words. For anyone building small automations, that is the kind of change you feel in your bill before you feel it in a benchmark.
Who Can Actually Use It
Neon is rolling out in the Gemini API in Google AI Studio, in Gemini for business, in Google Antigravity for paid Pro and Ultra subscribers, and in the Gemini app for Pro and Ultra subscribers. There is no free tier. If you are on a free plan, this launch does not change anything for you yet.
So here is the question I keep coming back to. If the middle model handles most of the hard work for less money, how often will you still reach for the biggest one?

