NEWS & VIEWS · AI MODELS

Google launches Gemini 4 Neon, a faster, cheaper model for AI Pro and Ultra subscribers

Neon sits between Gemini 4 Carbon and Hydrogen. Who can pick it today, what it costs developers, and what Google says it does better.

Google's launch image: an orange neon sign reading Gemini 4 Neon with the Gemini spark, glowing on a dark wall.
Image: Google

TL;DR

What is Gemini 4 Neon?

Google released Gemini 4 Neon on November 18, 2026. It sits between Gemini 4 Carbon, Google's model for the hardest work, and the fast everyday Hydrogen. It's rolling out now to Google AI Pro and Ultra subscribers in the Gemini app. If you pay for one of those plans, look for it in the model menu.

What did Google release?

On Wednesday, November 18, 2026, Google introduced Gemini 4 Neon. Google CEO Sundar Pichai posted about it on X at 4:33pm ET, and the model began rolling out the same day, so not everyone has it yet.

Google describes Neon as "frontier-level intelligence," meaning close to the most capable models, with fast end-to-end task completion for coding, office work and science. In the Gemini 4 family, Carbon handles your hardest, long-horizon projects and Hydrogen handles fast everyday jobs. Neon balances the two. It arrives a day after OpenAI released GPT-6.2 Astra and Sol.

What does Neon do better, according to Google?

  • Writing and documents. Google says Neon writes clearly with a natural, direct tone, reads uploads like spreadsheets, PDFs and photos more precisely, and turns dense material into reports and summaries. It can pull in details from Google apps like Gmail and Docs.
  • Research. It often brings in details from across the web to keep answers current, for jobs like comparing vendor prices before a big purchase.
  • Tasks in Gemini. Tasks, where Gemini works through multi-step jobs for you, can now run on Neon. Google says that gets complex jobs done more quickly than larger models while "stretching usage limits further."
  • Fewer made-up answers. Google says Neon recorded its lowest fabrication rate yet in internal Google Antigravity checks.

Who can use Gemini 4 Neon?

Source: Google's launch post and the @GeminiApp post, checked November 18, 2026. Google doesn't mention free Gemini app users.
WhereWho gets it
Gemini app (mobile, desktop and web)Google AI Pro and Ultra subscribers, from the model drop-down
Gemini API in Google AI Studio and Android StudioDevelopers on the paid tier
Google AI StudioAll Pro and Ultra subscribers can try it
Google AntigravityPaid Google AI Pro and Ultra subscribers
Gemini for businessVia the API or the model drop-down
Neon for cyberVetted defenders in Google's Fairwind Program, over the next few days

What does Neon cost in the API?

From Google's launch post, price checked Nov 18, 2026.
Gemini 4 NeonPrice per million tokens
InputUS$3
Cached inputUS$0.15 (a 95% discount)
OutputUS$15

Neon has a 1 million-token context window, the amount of text it can take in at once, and four thinking levels so teams can trade depth for cost. Tokens are the chunks of text AI models read and write.

How do the benchmarks look?

These are results Google chose to highlight, many from outside leaderboards:

  • Artificial Analysis Intelligence Index: #3 overall with a score of 61 at its highest "xhigh" thinking level, just behind larger, slower models.
  • Vals Index: #2, at 70.9%.
  • Artificial Analysis Coding Agent Index: #1, at 58.6.
  • Zapier AutomationBench and Andon Labs' Vending-Bench 2: #1 on both, at 56.3% and US$16,843 in a simulated business.
  • Cost: at its default "high" level, Google says Neon costs 32% less per task than Carbon, and roughly 50% to 80% less per task than top competing frontier models.

The jargon, in one line each

TermWhat it means
Frontier-levelClose to the most capable AI models available
Thinking levelsSettings that let the model reason longer for harder jobs, at higher cost
Context windowHow much text, files or data a model can work with at once
Cached inputText the API has already processed and can reuse for less
Google AntigravityGoogle's coding tool where AI agents build and edit software

X reactions

Google's Gemini app account confirmed who gets Neon, and benchmark group Vals AI pointed to a change in how it researches.

Our take

If you pay for Google AI Pro or Ultra, Neon looks like the sensible middle pick: try it on the reports, spreadsheets and price comparisons you'd normally hand to Carbon, and keep Carbon for the truly hard jobs. Running Tasks on Neon could also make your plan's limits last longer, if Google's claims hold.

The rankings are Google's own selection, and two AI labs shipped new models two days running. Judge Neon on your own files before you change plans or tools.

— Artificially Nerdy

What to do this week

  1. On Google AI Pro or Ultra? Open the Gemini app and check the model drop-down. If Neon isn't there yet, it's still rolling out.
  2. Try it on one real document, like a spreadsheet you need summarized or a vendor price comparison.
  3. Use Tasks in Gemini? Run your next multi-step task on Neon and see if it finishes faster.
  4. Build on the Gemini API? Price your workload at US$3 in and US$15 out per million tokens, and reuse prompt text to get the cached rate.

FAQ

Is Gemini 4 Neon free?

Google hasn't said so. In the Gemini app, it's rolling out to Google AI Pro and Ultra subscribers, and in the API it's on the paid tier.

Is Gemini 4 Neon better than Gemini 4 Carbon?

No. Google positions Carbon for the hardest, long-horizon projects and Neon as the faster, cheaper middle option. Google says Neon costs 32% less per task than Carbon at its default setting.

How much does Gemini 4 Neon cost in the API?

US$3 per million input tokens, US$0.15 per million cached input tokens and US$15 per million output tokens, with a 1 million-token context window.

Why don't I see Neon in my Gemini app?

Google says it's rolling out starting November 18, 2026, so it may take time to reach every Pro and Ultra account.

Sources

  1. Google: Introducing Gemini 4 Neon — Koray Kavukcuoglu, November 18, 2026. Availability, API price, context window, thinking levels, benchmarks, cost per task, Gemini app features, Tasks on Neon, Fairwind Program.
  2. Google Gemini (@GeminiApp) on X — November 18, 2026, 4:37pm ET.
  3. Sundar Pichai (@sundarpichai) on X — November 18, 2026, 4:33pm ET.
  4. Google DeepMind (@GoogleDeepMind) on X — November 18, 2026, 4:34pm ET. Between Carbon and Hydrogen, rolling out now.
  5. Vals AI (@ValsAI) on X — November 18, 2026, 4:35pm ET.

THE NERD SIGNAL

Rocky digs through the AI noise. You get the good bits.