Google has launched Gemini 4 Neon, a model they say brings frontier-level intelligence to the realms of complex coding, knowledge work, and science work without the long wait. If you’re paying for Google AI Pro or Ultra, you can now select Neon from the model menu in the Gemini app too. When I went into the app, it was immediately apparent that Neon has taken the place of Argon in the model picker. It’s now under the ‘Deeper analysis’ heading, with Hydrogen under ‘Everyday help’.
Neon is also being rolled out in the paid Gemini API in Google AI Studio and Android Studio, for Google Antigravity Pro and Ultra subscribers and in Gemini for business. The model is listed in the Vertex AI documentation for verified enterprise customers in the global, US, and EU regions.
The naming scheme for these new models is a complete mess, so let’s break that down first. Neon sits between Gemini 4 Carbon, which Google keeps for the most difficult long-horizon projects, and Gemini 4 Hydrogen, the faster, more everyday model. API pricing is $3 per million input tokens and $15 per million output tokens (with cached input at $0.15). The context window is 1 million tokens. It has four thinking levels, with ‘high’ as the default.
“To me Neon feels like the right mix of intelligence, speed and usability,” wrote Logan Kilpatrick in a post. “It’s absolutely punching well above its model class in a meaningful way.”
How Google says Neon compares
Google’s own comparison table puts Neon ahead of both Claude Opus 5.5 and OpenAI’s GPT-6.2 Sol on the Artificial Analysis Intelligence Index (score 61 v 59 v 57). Google states that 61 puts Neon in third place overall, behind the bigger, slower models. Neon comes out on top on the Vals Index (70.9%) and on AutomationBench (56.3%). It is behind Opus 5.5 on Terminal-Bench 4.1 and FrontierCode 2.0, and behind GPT-6.2 Sol on Chartography.

At its default ‘high thinking’ level, Neon is effectively on par with Carbon on the same index, at 32% lower cost per task and 50-80% less per task than the leading frontier models.
What it means if you already pay for Gemini
In the Gemini app, Google says Neon is for writing, for research, and for all the messy day-to-day stuff like reading a pile of receipts or checking your utility bill. It can work across spreadsheets, PDFs, charts, audio, and video all in one prompt. Those are Google’s own benchmarks, so independent tests may well come out differently.
It also connects seamlessly with Gemini’s Tasks, drawing context from Gmail and Docs to enhance your experience beyond what you get with heavier models like Carbon. Plus, Pro and Ultra subscribers can enjoy Neon in Google Antigravity, featuring a handy Orchestrator mode that combines Neon’s planning with Hydrogen’s speedy execution, along with access to Google AI Studio!
