- Google says Gemini 4 Neon scores 61 on the Artificial Analysis Intelligence Index at its xhigh thinking level, third overall, and ranks first on the Artificial Analysis Coding Agent Index at 58.6.
- Neon costs $3 per million input tokens and $15 per million output tokens, with a 1 million token context window and cached input at $0.15.
- Unlike Argon and Carbon, which went to vetted cyber defenders first, Neon is available on day one in the Gemini API, Google Antigravity and the Gemini app for paid subscribers.
Google released Gemini 4 Neon on November 18, a mid-tier model it prices at $3 per million input tokens and $15 per million output tokens, and says it ranks third overall on the Artificial Analysis Intelligence Index while running faster and cheaper than the larger models ahead of it.
Gemini 4 Neon fills the gap between Hydrogen and Carbon
Neon is the fourth Gemini 4 model in seven weeks. Google DeepMind's Koray Kavukcuoglu, who wrote the launch post, places it between Gemini 4 Carbon, the long-horizon flagship, and Gemini 4 Hydrogen, the fast model that now runs Google Search. Neon targets complex coding, analysis and enterprise workflows where Carbon is slower and more expensive than the task needs.
The benchmark claims are broad. Google reports a score of 61 on the Artificial Analysis Intelligence Index at xhigh, second place on the Vals Index at 70.9%, 65.2% on Terminal-Bench 4.1 against Hydrogen's 57%, and second place on Arena's new Agent Arena Index with a score of 1336 across more than 4,200 real agent sessions. All of these figures are self-reported by Google.
Neon's $15 output price undercuts Carbon's standard rate
Pricing is where Neon earns its slot. At $3 and $15 per million tokens, Neon costs twice as much as Hydrogen and 25% less than Carbon's standard rate of $4 and $20, which applies once Carbon's introductory period ends. Google says that at its default high thinking level Neon has a 32% lower cost per task than Carbon, and that at xhigh it still costs at least 36% less per task than competing models at their highest settings.
| Gemini 4 Neon at a glance | |
|---|---|
| Released | November 18, 2026 |
| API price | $3 input, $15 output per million tokens ($0.15 cached input) |
| Context window | 1 million tokens |
| Thinking levels | Four, up to xhigh |
| Artificial Analysis Intelligence Index | 61 at xhigh, third overall |
| Terminal-Bench 4.1 | 65.2% |
| SEC-bench Pro (vulnerability finding) | 90.5% |
| Availability | Gemini API, Gemini for business, Google Antigravity, Gemini app (Pro and Ultra) |
"While Carbon tackles your hardest, long-horizon projects and Hydrogen handles fast everyday execution, Neon balances depth and efficiency."
Koray Kavukcuoglu, SVP, Google DeepMind, Introducing Gemini 4 Neon, November 18, 2026
The rollout breaks Google's recent pattern. Argon and Carbon launched to Fairwind Program cyber defenders before anyone else. Neon goes to paying developers, enterprises and Gemini app subscribers on day one, while a cyber-tuned version reaches Fairwind members over the coming days. Google pairs that wider release with real-time misuse safeguards it describes as comparable to Carbon's, and reports that Neon refused 1.98% of Terminal-Bench 4.0 tasks.
Google now sells four Gemini 4 models at four price points, and the one it put in front of every paying developer first is the middle one.
Neon gives Google a mid-priced model for long-running AI agent work, with a context window of 1 million tokens and the top score on the Artificial Analysis Coding Agent Index by Google's account. Its $15 output rate is a direct bid for developers who find Carbon too costly and Hydrogen too light, and with general availability on launch day, the benchmark claims can be tested by customers within weeks.
Santage is committed to independent, transparent journalism. This article is produced in accordance with Santage's Editorial Standards and aims to provide accurate and timely information. Readers are encouraged to verify information independently.