← Shelf service.News

Gemini 4 Neon, the ranking and the price catch

Google introduced Gemini 4 Neon on 18 November 2026, the newest model of the Gemini 4 family after Argon, Carbon and Hydrogen. Here is what the reel said, with a source next to every number.

Gemini 4 Neon key art: the name as an orange neon sign on a dark wall.
Google, launch postOpen source

Why it matters

If you build with coding agents, Neon in Google's Antigravity CLI now tops the Artificial Analysis Coding Agent Index, one point ahead of Claude Code. If you pick a Gemini model by price, Carbon is the cheaper one until 30 November, and from 1 December it is Neon.

My read: it is too early to call the race between ChatGPT, Claude and Google, so keep an open mind and try it.

What it is

What it is

Google's Gemini 4 family table: Carbon, Neon and Hydrogen by use, quality, price, thinking levels, speed, plans and availability.
Google, launch postSource

Neon is the middle model of the Gemini 4 family. In Google's words, Carbon tackles your hardest, long-horizon projects, Hydrogen handles fast everyday execution, and Neon balances depth and efficiency. Koray Kavukcuoglu, SVP at Google DeepMind, announced it on 18 November.

At its default high thinking level, Google says Neon costs 32 % less per task than Carbon. It has a context window of 1 million tokens.

Sources: Google's launch post and Gemini 4 Neon model card (PDF)

Where it ranks

Where it ranks

Google's benchmark table: Gemini 4 Neon against Claude Opus 5.5 and GPT-6.2 Sol across coding, knowledge work, science, reasoning, multimodal, long context and security.
Google, launch postSource

On the Artificial Analysis Coding Agent Index, Antigravity CLI with Gemini 4 Neon (xhigh) scores 59. Claude Code with Claude Fable 5.2 (xhigh) and Claude Code with Claude Opus 5.5 (max) follow at 58. Codex with GPT-6.2 Astra (max) scores 53. Google's post gives Neon 58.6 on the same index.

On the Artificial Analysis Intelligence Index, Google says Neon ranks third overall with 61 at xhigh, just behind larger, slower models.

In Google's own table, Neon leads Claude Opus 5.5 and GPT-6.2 Sol on most of the benchmarks Google picked. Opus 5.5 is ahead on Terminal-bench 4.1 and FrontierCode 2.0, and GPT-6.2 Sol on Chartography.

Sources: Artificial Analysis, coding agents and Google's launch post

What it costs

What it costs

Google's chart: Artificial Analysis Intelligence Index against cost per task for Gemini 4 Neon, Gemini 4 Carbon, Claude Opus 5.5 and GPT-6.2 Sol.
Google, launch postSource

API prices in USD per million tokens. Until 30 November Carbon is the cheaper of the two. From 1 December it is Neon.

What it costs
ModelInputOutput
Gemini 4 Neon315
Gemini 4 Hydrogen1.507.50
Gemini 4 Carbon, introductory price until 30 November 2026210
Gemini 4 Carbon, from 1 December 2026420

Cached input for Neon is USD 0.15 per million tokens. Carbon has been on the API for paying developers since 10 November.

Sources: Google, Neon and Google, Carbon

Who has it

Who has it

Google says Neon is rolling out, so it can take a little while to show up. A version for cyber security goes to trusted defenders in Google's Fairwind Program.

  • Developers, through the Gemini API in AI Studio (paid tier).
  • Businesses, through Gemini for business.
  • Paid Google AI Pro and Ultra subscribers, in Google Antigravity.
  • Google AI Pro and Ultra subscribers, in the Gemini app, from the model drop-down.

Source: Google's launch post

The Club, once a week

Your feed.
With an editor.

What moved, what matters and what I’m keeping an eye on. Plus the news, verdicts and short lists worth taking with you.

Put me on the list
Next in News

Google Maps Platform comes to Google AI Studio

Google says you can now use Maps Platform inside AI Studio without leaving the workspace...

Read the News

Full library