Key points
- How Google prices Gemini 4 Neon against Carbon and Hydrogen
- Why Google's Carbon costs less than Neon until Dec. 1
- Which Google AI plans include Neon, Carbon, and Hydrogen
Google (GOOGL) released Gemini 4 Neon on Nov. 18, a model that sits between Gemini 4 Carbon, which Google calls its most capable reasoning model, and the faster and cheaper Gemini 4 Hydrogen. Neon costs $3 per million input tokens and $15 per million output tokens in the Gemini API, twice Hydrogen's price.
Through Nov. 30, Neon also costs more per token than Carbon. Google's API pricing page lists Carbon at an introductory $2 per million input tokens and $10 per million output tokens through Nov. 30, rising to $4 and $20 on Dec. 1. From then on, Neon will cost 25% less per token than Carbon.
In the Gemini app, Neon is included in the $19.99-a-month Google AI Pro plan, while Carbon requires the Ultra plan, which starts at $99.99 a month.
Neon costs twice as much as Hydrogen per token
All three models have a 1 million-token context window. Neon's caching discount cuts the price of repeated input to $0.15 per million tokens, a 95% discount, according to Google's announcement.
| Model | Input, per 1M tokens | Output, per 1M tokens | Cached input, per 1M tokens | Free API tier |
|---|---|---|---|---|
| Gemini 4 Carbon | $2.00 through Nov. 30; $4.00 from Dec. 1 | $10.00 through Nov. 30; $20.00 from Dec. 1 | $0.10 through Nov. 30; $0.20 from Dec. 1 | No |
| Gemini 4 Neon | $3.00 | $15.00 | $0.15 | No |
| Gemini 4 Hydrogen | $1.50 | $7.50 | $0.075 | Yes |
| Gemini 4 Argon | $2.00 through Nov. 30; $4.00 from Dec. 1 | $10.00 through Nov. 30; $20.00 from Dec. 1 | $0.10 through Nov. 30; $0.20 from Dec. 1 | No |
Standard paid-tier prices from Google's Gemini API pricing page as of Nov. 19, 2026. Output prices include thinking tokens.
Gemini 4 Argon, which Google opened to developers in October, is still listed in the API at the same prices as Carbon. Google's developer guide for Neon includes instructions for moving to it from Argon and the other Gemini 4 models, and from Gemini 3.8 Flash and earlier Gemini 3 models. Google's plan page for the Gemini app lists Neon, Carbon, and Hydrogen but not Argon.
Neon ranks one point ahead of Carbon on Artificial Analysis
Google said Neon ranks third overall on the Artificial Analysis Intelligence Index, with a score of 61 at its highest thinking setting. The independent benchmarking firm's leaderboard showed the same score on Nov. 19, behind Anthropic's Claude Fable 5.2 at 65 and OpenAI's GPT-6.2 Astra at 62.
| Model | Setting tested | Intelligence Index | Cost per task | Time per task |
|---|---|---|---|---|
| Gemini 4 Neon | Xhigh | 61 | $6.44 | 21.5 minutes |
| Gemini 4 Carbon | High | 60 | $7.76 | 36.6 minutes |
| Claude Opus 5.5 | Max | 59 | $10.04 | 25.1 minutes |
| Gemini 4 Hydrogen | Xhigh | 54 | $3.16 | 15.2 minutes |
| Gemini 4 Argon | High | 50 | $4.97 | 29.5 minutes |
Source: Artificial Analysis leaderboard as of 8:55 a.m. ET, Nov. 19, 2026. Each model is listed at the thinking setting Artificial Analysis tested.
Artificial Analysis tested both Google models at their top settings: xhigh for Neon and high for Carbon, whose settings stop at high. On that basis, Neon finished its tasks in about 59% of Carbon's time and cost about 17% less per task. Google's own comparison, at Neon's default high setting, puts Neon's cost per task 32% below Carbon's.
Neon's default thinking level is high, and its other levels are low, medium, and xhigh. Higher levels spend more tokens on reasoning, and those tokens are billed as output.
Google says Neon is built for coding and analysis
Google describes Carbon as the model for "your hardest, long-horizon projects" and Hydrogen as the one for "fast everyday execution," with Neon in between. The company says Neon ranks first on the Artificial Analysis Coding Agent Index at 58.6 and scores 65.2% on Terminal-Bench 4.1. On the Vals Index, Google says Neon ranks second at 70.9%.
In Google Antigravity, the company's coding platform, Neon can direct several Hydrogen agents working in parallel. A version tuned for cybersecurity work, which Google calls Neon for cyber, is rolling out over the next few days to defenders in Google's Fairwind Program.
Neon costs 25% less per token than Claude Opus 5.5
Anthropic charges $4 per million input tokens and $20 per million output tokens for Claude Opus 5.5, according to its pricing page, so Neon's token prices are 25% lower. On Artificial Analysis, Neon scored 2 points higher than Opus 5.5 and cost $6.44 per task against $10.04.
Google's Pro plan includes Neon, and Ultra adds Carbon
In the Gemini app, the $4.99-a-month Google AI Plus plan includes Hydrogen, according to Google's plan page. The free plan lists automatic model selection without naming a Gemini 4 model. Google AI Pro, at $19.99 a month, adds Neon. Google AI Ultra, at $99.99 or $199.99 a month, adds Carbon and gives the highest usage limits for all three models.
| Plan | Monthly price | Gemini 4 models in the app |
|---|---|---|
| Free | $0 | Automatic model selection |
| Google AI Plus | $4.99 | Hydrogen |
| Google AI Pro | $19.99 | Neon, Hydrogen |
| Google AI Ultra | $99.99 or $199.99 | Carbon, Neon, Hydrogen |
US prices from Google's plan page as of Nov. 19, 2026.












