SIAgentNews.com
← All news
DevelopingGemini

Gemini 4 Neon is already on the $19.99 Pro plan, so Ultra is not required

Google's own plan pages put Gemini 4 Neon on the $19.99 AI Pro tier. The $99.99 Ultra tier mostly buys a bigger usage allowance and access to a different model.

· First reported by OrcaRouter

$19.99monthly price of the Google AI Pro plan that includes Neon

The short version

  • Neon has been included on Google AI Pro ($19.99) since general availability on 18 November 2026.
  • Ultra ($99.99) gives a larger Neon allowance and adds Gemini 4 Carbon, which Pro lacks.
  • Usage is metered by compute, not message count, refreshing every five hours under a weekly cap.
  • Neon's context window is 1M tokens in the app and API, but up to 512K in Antigravity.
  • Ultra's 20x tier shows no price on Google's plan page, so the figures here cover the 5x tier.

Why it matters

If you are weighing an $80 monthly jump to Ultra because of Neon, the model is already included in the cheaper plan. The real question is whether you keep running out of your Pro allowance, or need Carbon.

A testimonial that muddied the picture

On 20 November, the account @kimmonismus (Chubby) posted a screenshot saying Gemini 4 Neon was excellent and that he had upgraded to the Ultra tier because of it. The post collected 1,216 likes, and replies quickly asked whether Neon requires Ultra's $99.99 price.

According to OrcaRouter, which compared Google's plan and product pages on 21 November 2026, it does not. Neon has been part of the $19.99 Google AI Pro plan since it became generally available on 18 November 2026. It also comes with Ultra and with Gemini Enterprise.

This fits what our hosts told you today: a reported detail from a third-party outlet, so check Google's own plan page before spending anything.

What each plan actually includes

Google sells three Gemini 4 model tiers in its consumer products. The $9.99 AI Plus plan reaches only the smallest, called Hydrogen. Pro adds Neon. Ultra includes all three, adding Gemini 4 Carbon, the strongest reasoning model in Google's consumer lineup.

The confusing part is the wording. Pro's listing says it has access to Neon, while Ultra's says higher access. That one word describes a bigger quantity, not a new permission.

Google's Antigravity coding tool, which is Google's workspace for AI-assisted coding, confirms this in a table. Neon is unavailable on Free and Plus, and available on Pro, Ultra and Enterprise. It also sits at the top of the model picker and is the default for new conversations, so a Pro subscriber is likely using it already.

  • Carbon in the Gemini app and in Antigravity: Ultra only.
  • Neon in both: Pro, Ultra and Enterprise.
  • Hydrogen: Plus, Pro and Ultra.

What the extra $80 buys

The gap between Pro and the 5x Ultra tier is $80 a month ($99.99 versus $19.99). The page showed no annual option for Ultra. Google does not show a price for the 20x tier either; that column's price cell is empty, so the numbers here cover only the 5x tier.

The difference is quota. Google measures usage by compute, a mix of how complex your prompt is, which models and features you use, and how long the chat runs. Allowances refresh every five hours and stop at a weekly cap. The tiers are described relative to each other: Plus gets double the free allowance, Pro gets four times it, and Ultra gets five or twenty times Pro's, depending on the tier.

Because Google counts compute rather than messages, nobody can honestly translate $80 into a number of extra prompts. For comparison, OrcaRouter notes that Neon's developer pricing is $3.00 per million input tokens and $15.00 per million output tokens. At those rates $80 covers 5.33 million output tokens or 26.7 million input tokens. A token is a small chunk of text, roughly a word fragment.

Pro and Ultra subscribers can also buy extra credits past the cap, controlled by a setting with two values, Never and Always.

One model, two context limits

A context window is how much text a model can consider at once. Google's developer changelog describes Neon as having a 1M-token window, and the Gemini app lists 1 million for both Pro and Ultra. Antigravity's documentation says up to 512K tokens.

That ceiling depends on where you use the model, not on your plan. A Pro and an Ultra subscriber running the same agent in Antigravity get the same limit. If you plan a long job on a large codebase, this is the detail that could silently cut your input short.

Antigravity also tracks quota by model family. Gemini models have their own allowance, separate from the Claude and GPT models the tool offers. Time spent on those other models does not use up your Neon allowance.

How to decide

OrcaRouter's rule is simple: upgrade only if you can point to a specific week when you hit Pro's cap on Gemini models and still had work left. Otherwise you are paying for headroom you may never use. Ultra also makes sense if Carbon is what you want inside the app or Antigravity.

As for the testimonial, it is one screenshot with no benchmark or method. We cannot tell whether the $80 went to Neon, Carbon or other Ultra extras. For an independent figure, OrcaRouter points to Artificial Analysis, whose Intelligence Index scores Neon at 61.4 on its highest thinking setting, sixth of 96 models. At its default setting Neon scores 59.6, against 60.0 for Carbon at its maximum.

Programmatic work has a third route: pay per token through the API, with no subscription and no five-hour clock. OrcaRouter, a model-routing service, says it does not carry Neon in its catalogue, so that point is not a sales pitch for Neon.

What to watch

  • Whether Google clarifies the 1M versus 512K context difference between its app, API and Antigravity.
  • Whether Google publishes a price for the Ultra 20x tier on its plan page.
  • Whether Google rewords the Pro and Ultra listings so the allowance difference is clearer.

Try it

Open Google's plan comparison page and your Gemini model picker. Confirm Neon appears on your current tier, and note any week you hit the usage cap before paying for Ultra.

Sources and links

Read the original at OrcaRouter

First reported by OrcaRouter. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.

Liked it? Say why

Add a line for other readers.

No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.

More AI news