The short version: Gemini 4 Neon, released by Google on November 18, costs $3 per million input tokens and $15 per million output tokens. Today that is 50 percent more per token than Gemini 4 Argon and Gemini 4 Carbon, which sit at $2 and $10. On December 1, Google’s own pricing page doubles Argon and Carbon to $4 and $20. On December 1, Neon becomes the cheaper of Google’s big models without changing its own price at all: 25 percent cheaper per token than either sibling. If your business runs anything on the Gemini API, or pays a developer or app vendor who does, the model choice you made last month is about to be the wrong one on price.
That is the news in short. Below is the whole Gemini 4 price ladder before and after the two dates that reshape it, a worked monthly bill for a landscaping company’s estimates, what Google’s “32 percent cheaper per task” claim can and cannot mean this month, and where Neon actually shows up if you never touch an API. Every price here was read off Google’s Gemini API pricing page and its Gemini subscriptions page on November 19, 2026.
What is Gemini 4 Neon?
Neon is the fourth Gemini 4 model Google has shipped this fall, after Argon, Carbon and Hydrogen. In its launch post, Google leans heavily on coding and agent benchmarks, and describes a model with a one million token context window and four thinking levels, defaulting to “high.” Google reports it at second place on the Vals Index with 70.9 percent and first on the Artificial Analysis Coding Agent Index.
Benchmarks are Google’s numbers, and we have not run our own. What matters more for a small business is that Neon arrives at a price point that did not exist in the lineup before: halfway between the budget models and the big ones today, and below the big ones from December 1.
Gemini 4 Neon pricing against the rest of the lineup
Google’s pricing page lists two scheduled changes. Argon and Carbon move from $2 and $10 to $4 and $20 “starting December 1, 2026.” Gemini 3.8 Flash moves from $0.75 and $3.75 to $1.50 and $7.50 “starting January 1, 2027.” Neon and Hydrogen carry no scheduled change. Here is the ladder, per million tokens, standard tier:
| Model | Input today | Output today | Input, Jan. 2027 | Output, Jan. 2027 | Free API tier |
|---|---|---|---|---|---|
| Gemini 3.8 Flash | $0.75 | $3.75 | $1.50 | $7.50 | Yes |
| Gemini 4 Hydrogen | $1.50 | $7.50 | $1.50 | $7.50 | Yes |
| Gemini 4 Neon | $3.00 | $15.00 | $3.00 | $15.00 | No |
| Gemini 4 Argon | $2.00 | $10.00 | $4.00 | $20.00 | No |
| Gemini 4 Carbon | $2.00 | $10.00 | $4.00 | $20.00 | No |
Read the table from the bottom up and the story is plain. Since launch, Argon and Carbon have been sold on a cheaper dated rate, and that rate ends in 11 days. Neon was launched straight at its permanent price, which is why it looks expensive this week and reasonable next month. At the other end, Flash catches up with Hydrogen on January 1, so the budget tier collapses into a single price: $1.50 and $7.50. We covered that second change, and Hydrogen’s odd free-plan access, in our Gemini 4 Hydrogen post. The December 1 doubling for Argon was first dated in our Gemini 4 Carbon post, which also explains who can actually get Carbon.
What would a real small-business job cost on each model?
Take a seven-person landscaping company that uses a quoting tool built on the Gemini API. Each month it drafts 400 estimates and follow-up emails. Each job feeds the model about 3,000 tokens of price sheet, job notes and customer history, and gets back about 2,000 tokens, counting the model’s thinking, which Google bills as output. That is 1.2 million input tokens and 0.8 million output tokens a month.
- Gemini 3.8 Flash: $3.90 a month today, $7.80 from January 1.
- Gemini 4 Hydrogen: $7.80 a month, before and after.
- Gemini 4 Neon: $15.60 a month, before and after.
- Gemini 4 Argon or Carbon: $10.40 a month today, $20.80 from December 1.
Two things jump out. First, on the same token counts, a business sitting on Argon pays $5.20 a month more from December 1 than it would on Neon, a 25 percent saving for changing one setting. Second, the dollars are small. Nobody’s landscaping business lives or dies on $20 a month. The reason to care is that these are per-job prices, and they scale: a shop that grows from 400 to 4,000 jobs, or adds an agent that reads every inbound email, multiplies every line above by ten.
There is a cheaper Neon, too. The pricing page lists Neon’s batch and flex rates at $1.50 and $7.50, half of standard, for work that does not need an answer right away. Overnight estimate drafts or a weekly review summary fit that description. Run the same landscaping workload through batch and Neon costs $7.80, exactly what Hydrogen costs at full speed.
The honest caveat on all of this: our bill assumes every model uses the same number of tokens for the same job. They do not. A model that thinks longer bills more output, and a smarter model may finish in fewer tries. Price per token is where the comparison starts, not where it ends.
Is Neon really “32 percent cheaper per task” than Carbon?
Google’s launch post says that at its default high thinking level, Neon pairs a strong Artificial Analysis Intelligence Index score “with a 32% lower cost per task than Carbon.” The chart behind it is labeled as cost per task on the AA Index v5.01. What the post does not say is which Carbon price the comparison used: the $2 and $10 Carbon costs today, or the $4 and $20 it costs from December 1. That one missing detail changes what the claim means for you this month.
Here is the arithmetic, assuming both models use a similar mix of input and output tokens:
- If Google used today’s Carbon price, Neon costs 1.5 times as much per token, so to come out 32 percent cheaper per task it must use only about 45 percent as many tokens. After December 1, when Neon becomes 25 percent cheaper per token, the same task on Neon would cost about a third of what it costs on Carbon.
- If Google used the December 1 price, Neon is 25 percent cheaper per token and uses about 91 percent as many tokens. In that reading, the 32 percent saving holds from December 1, but this month, while Carbon is still on its introductory rate, Neon would cost about 36 percent more per task.
Either way, Neon wins on cost per task from December 1. What is genuinely unknown is whether it wins this month. The cheap fix for that uncertainty is not a better press release; it is your own usage log. Run 20 of your real jobs through each model in Google AI Studio and read the token counts. The model that is cheaper on your work is the one to pick, whatever any index says.
Can I use Gemini 4 Neon without touching the API?
Yes, with a subscription. Google’s launch post says Neon is “rolling out today” in the Gemini API in Google AI Studio, in Gemini for business through the API and the model drop-down, in Google Antigravity for paid Pro and Ultra subscribers, and in the Gemini app for Pro and Ultra subscribers.
On the subscriptions page, read November 19, the US prices are $19.99 a month for Google AI Pro, whose app features list “Access to Gemini 4 Neon,” and from $99.99 a month for Google AI Ultra, which lists “Highest access to Gemini 4 Neon.” The free plan lists Hydrogen and no Neon. One wrinkle: the $4.99 Google AI Plus card on the same page carries a line reading “Includes Gemini 4 Neon,” while the launch post names only Pro and Ultra for the app. If you pay for Plus, open the model picker and look before you upgrade on the strength of either page.
For a business owner who mostly writes, plans and answers email in the Gemini app, that is the practical route. Pro already covers Neon, so you are not paying per token at all; the price ladder above only matters to you if your software vendor or developer runs your workload through the API.
Does the free tier matter here?
It matters in one specific way. The pricing page lists Hydrogen and 3.8 Flash as free on the API’s free tier, and for that free tier the “Used to improve our products” row reads “Yes.” On the paid tier it reads “No.” Neon, Argon and Carbon have no free API tier at all, so anything you send them is paid and, per that same table, not used to improve Google’s products.
For a small business testing with customer names, addresses or quotes in the prompt, that is a reason to test on a paid key even when the free one would technically work. A few dollars buys you the privacy line you actually want.
What should a small business do before December 1?
Four steps, none of which needs a developer to understand:
- Find out which model your tools use. Ask your app vendor or developer one question: “Which Gemini model does our tool call, and on which tier?” If the answer is Argon or Carbon, your per-token cost doubles on December 1 unless something changes.
- Ask for a ten-job test. Have the same ten real jobs run on Neon and on the model you use now, and compare the output quality and the token count side by side. Quality first, price second.
- Move work that can wait to batch. Anything that does not need an instant reply, such as overnight drafts or weekly summaries, can run at half price on Neon’s batch rate.
- Put a spend cap on the key. Price changes are the moment a bill surprises people. A monthly cap on your API key turns a surprise into a pause.
The good news inside a price increase is that it forces a decision most owners never make: which model does this job need? Google now sells five sensible answers, from $1.50 to $4 per million input tokens, and the cheapest good answer for most small-business writing and quoting work is likely to be either Hydrogen or Neon from December 1. Picking on purpose, with your own ten jobs as the test, is the habit that pays for itself every time a vendor moves a price.
Frequently Asked Questions
How much does Gemini 4 Neon cost?
Google’s Gemini API pricing page, read November 19, 2026, lists Gemini 4 Neon at $3 per million input tokens and $15 per million output tokens on the standard tier, with cached input at $0.15 per million and batch or flex processing at half the standard rate. In the Gemini app, Neon comes with the $19.99 a month Google AI Pro plan and the Ultra plans.
Is Gemini 4 Neon cheaper than Gemini 4 Argon and Carbon?
Not this month, but from December 1, 2026, yes. Argon and Carbon cost $2 and $10 per million tokens until November 30 and $4 and $20 from December 1, while Neon stays at $3 and $15, which makes it 25 percent cheaper per token than either from that date.
Is Gemini 4 Neon available on the free tier?
No. Google’s pricing page says Gemini 4 Neon is not available on the free API tier, and the Gemini app’s free plan lists Gemini 4 Hydrogen rather than Neon. Neon is available to Google AI Pro and Ultra subscribers in the app and to paid API users.
Which Gemini model should a small business use for everyday writing?
For drafting emails, estimates and summaries, start by testing Gemini 4 Hydrogen and Gemini 4 Neon on ten of your real jobs and comparing quality and token counts. From January 1, 2027, Hydrogen and Gemini 3.8 Flash share the lowest price at $1.50 and $7.50 per million tokens, while Neon is the cheapest of the larger models from December 1.
Which Gemini model is your business running on right now, and do you know what it will cost you on December 1? Tell us in the comments.
