Gemini 4 Neon: What It Is, Pricing and Who Can Use It
Google released Gemini 4 Neon on November 18, 2026, a model that sits between Carbon, its most capable, and Hydrogen, its fast one. It is rolling out to the paid Gemini API, Gemini for business, Antigravity and the Gemini app for AI Pro and Ultra subscribers. In the API it costs $3 per million input tokens.
What we know
- What it is, according to Google: a model that balances depth and efficiency, aimed at complex coding, analysis and production business workflows. Carbon stays the pick for the hardest long-running projects, and Hydrogen for quick everyday tasks.
- API pricing: $3 per million input tokens, $0.15 per million cached input tokens and $15 per million output tokens, per the announcement and the API pricing page. That page lists no free tier for Neon.
- Context and reasoning: up to one million tokens of context and four reasoning levels. The announcement names "high", the default, and "xhigh", where Google reports its best results on the hardest tasks.
- Who gets it: developers on the paid tier of the Gemini API (Google AI Studio and Android Studio), companies on Gemini for business, and Google AI Pro and Ultra subscribers in Antigravity and in the Gemini app, through the model picker. The announcement does not mention the free plan.
- Cost per task, according to Google: at the default level, Neon spends 32% less per task than Carbon. Per token, though, Neon costs more than Carbon until November 30 and less from December 1.
- Performance, on benchmarks Google picked: first on Artificial Analysis's Coding Agent Index (58.6) and third on its Intelligence Index (61 at xhigh). Google also says it recorded its lowest fabrication rate yet in internal Antigravity testing.
- Cyber defense: a version for trusted defenders arrives "in the coming days" through the Fairwind Program, Google says.
What changes and what doesn't
The price map changes. Using the API pricing page, a task with 30,000 input tokens and 8,000 output tokens costs about $0.21 on Neon, $0.105 on Hydrogen and $0.14 on Carbon until November 30. From December 1, Carbon moves to $4 and $20 per million and the same task rises to about $0.28. Per 1,000 tasks that is roughly $210, $105, $140 and $280. This is an estimate made for this article with the same token counts for every model: if Neon finishes the task with fewer tokens, its real cost drops.
What doesn't change: the performance figures come from benchmarks Google chose. And a new model can answer differently from the one you run today, so switching models in production needs your own tests.
How to tell if it affects you
- Check which model your site calls. If your chatbot, site search or product-copy generator uses the Gemini API, look for the model name in its settings (for example gemini-4-carbon or gemini-4-hydrogen).
- Run the December math. If you use Carbon or Argon, your per-token price doubles on December 1 even if you change nothing. That is the moment to compare against Neon.
- Test with 20 real cases. Send the same tasks to your current model and to Neon, and compare quality, tokens spent and time. If Hydrogen already does the job, there is no need to move up.
- Start at the default level. Save maximum reasoning for work that truly needs it: reasoning tokens are billed as output.
- Review what it publishes. If the model writes copy or code for your site, review it before it goes live: a lower fabrication rate in Google's tests is not zero.
Related: Gemini 4 Carbon Now Available: Who Can Use It, Pricing
Sources
- Google, "Introducing Gemini 4 Neon," Koray Kavukcuoglu, November 18, 2026
- Google AI Studio, Gemini API pricing
- Google AI Developers Forum, "Gemini 4 Neon and Orchestrator (Neon) are now available in Google Antigravity," November 18, 2026
Updates: if Google opens Neon to the free plan, changes its pricing or names all four reasoning levels, it will be added here with a link.