
On this page
Short answer
Google has released Gemini 4 Neon, a new model that sits between the fast Gemini 4 Hydrogen and the most powerful Gemini 4 Carbon. Google says it brings frontier-level reasoning with faster, more efficient task completion. Since November 18 it is rolling out in the Gemini app for Google AI Pro and Ultra subscribers, where it appears as "Deeper analysis" in the model picker, and for developers in the Gemini API and Google Antigravity.
What happened
Google announced Neon on its official blog on November 18, 2026, in a post by Koray Kavukcuoglu, Chief AI Architect at Google. The key claims, as stated by Google:
- Ranked #3 overall with a score of 61 on the Artificial Analysis Intelligence Index at its highest thinking level, just behind larger, slower models. Artificial Analysis confirmed the score in its own write-up.
- #1 on the Artificial Analysis Coding Agent Index and #2 on the new Agent Arena Index, including #2 for steerability, meaning it adapts cleanly when you correct it mid-task.
- Its lowest fabrication rate yet in Google's internal testing, plus better multi-step instruction following.
- A 1 million-token context window and four thinking levels, so it can read long PDFs, spreadsheets, charts, audio and video in one go.
According to 9to5Google, Neon replaces the previous Argon model in the Gemini app. The picker now reads: Auto, Everyday help (4 Hydrogen), Deeper analysis (4 Neon) and Code & complex work (4 Carbon).
Why people are talking about it
Neon was one of the top stories of the day. Google's announcement reached the Hacker News front page with more than 700 points, and Techmeme linked coverage from 9to5Google, SiliconANGLE, VentureBeat, Digital Trends, Neowin and Artificial Analysis. A big reason is the surprise: SiliconANGLE notes that Neon, a mid-tier model, beat the flagship Carbon on some of Google's benchmarks. Most of these numbers come from Google's own tests, so independent results over the coming weeks will matter.
What it means if you use AI
- Better everyday writing and research. Google says Neon writes in a clear, natural, direct tone, reads uploaded files more precisely and pulls in fresh details from the web for fast-moving topics.
- Tasks get a stronger engine. Tasks in Gemini can now run on Neon, which Google says handles complex, multi-step jobs faster while stretching usage limits further.
- More honest answers. A lower fabrication rate means fewer made-up details, which matters whenever you rely on an AI for facts. It is still worth double-checking anything important.
Quick checklist
- Open the Gemini app and look for "Deeper analysis" in the model picker (AI Pro or Ultra plan).
- Try it on a long document or spreadsheet and ask for a short summary.
- Use the right tier: Hydrogen for quick questions, Neon for analysis, Carbon for the hardest projects.
- If you connect Gmail or Docs, review which apps Gemini can access in its settings.
- Developers can test Neon in Google AI Studio or Antigravity on the paid API tier.
Related reading
- Gemini 4 Hydrogen: 2.5x faster AI in Gemini and Search
- Gemini 4 Carbon: Google's new #1 AI model, coming soon
- Gemini Tasks goes free in 6 countries: what it can do
Sources
- Google, The Keyword (Nov 18, 2026): Introducing Gemini 4 Neon
- 9to5Google (Nov 18, 2026): Google launches Gemini 4 Neon with Carbon quality at Flash speeds, live for AI Pro
- SiliconANGLE (Nov 18, 2026): Google's latest midtier model Gemini 4 Neon is three times faster than Carbon and wins on some metrics
- Artificial Analysis (Nov 18, 2026): Gemini 4 Neon intelligence gain
Read on November 19, 2026.
Sources
- Google, The Keyword (Nov 18, 2026): Introducing Gemini 4 Neon
- 9to5Google (Nov 18, 2026): Google launches Gemini 4 Neon with Carbon quality at Flash speeds, live for AI Pro
- SiliconANGLE (Nov 18, 2026): Google's latest midtier model Gemini 4 Neon is three times faster than Carbon and wins on some metrics
- Artificial Analysis (Nov 18, 2026): Gemini 4 Neon intelligence gain
- Cover image: Google-Deep Mind headquarters in London, 6 Pancras Square, by Gciriani, Wikimedia Commons, CC BY-SA 4.0
Written by our editorial team with the help of AI tools. Spotted a mistake? Tell us in the discussion.