top of page

GitHub Copilot adds Gemini 4 Neon at $3/$15, but not on Copilot Pro

1 hour ago
6 min read
GitHub Copilot adds Gemini 4 Neon at $3/$15, but not on Copilot Pro, Data Studios original cover

GitHub made Gemini 4 Neon generally available in GitHub Copilot on November 18, 2026, the same day Google released the model in the Gemini API. The model appears in the Copilot model picker for subscribers on Copilot Pro+, Max, Business and Enterprise, and GitHub bills it at the provider's list price under usage-based billing. Copilot Free and Copilot Pro are absent from the list of eligible plans, which separates Neon from Gemini 4 Hydrogen, the cheaper Google model that GitHub opened to Pro users as well on November 13.


Google's official price list for Gemini 4 Neon is $3.00 per million input tokens, $0.15 per million cached input tokens and $15.00 per million output tokens, with thinking tokens billed as output. Because GitHub converts token usage into GitHub AI Credits at a fixed rate of $0.01 per credit, those list prices translate directly into credits drawn from a plan's monthly allowance, and into overage charges once the allowance runs out.


GitHub's changelog says the rollout is gradual, so the model may take some time to appear in every account. When it does, it is selectable in Visual Studio Code, Visual Studio, Copilot CLI, the Copilot coding agent, the Copilot app, github.com, GitHub Mobile on iOS and Android, JetBrains IDEs, Xcode and Eclipse.


··········


GEMINI 4 NEON IN GITHUB COPILOT: PLANS, CLIENTS AND ADMIN POLICY.


GitHub positions Gemini 4 Neon for agentic coding, multistep tool use and knowledge work. The changelog repeats Google's own claims for the model: quality close to Gemini 4 Carbon at a lower token cost per task, faster responses, fewer unrequested changes during coding tasks, and a lower tendency to make unsupported claims about code, tools or test results. GitHub did not publish its own evaluation of the model in the announcement, so these remain vendor statements.


For organizations, access follows the existing model policy. Business and Enterprise administrators manage Neon from the model policy in Copilot settings, and under default model enablement a newly added model is switched on automatically unless an administrator has turned off the global default or disabled this specific model. Companies that review every new model before developers can use it should check that setting now, because Neon will otherwise become available to their seats as the rollout reaches them.


........


Copilot plan

Monthly price

Monthly AI credits

Gemini 4 Neon

Gemini 4 Hydrogen

Copilot Free

Free

Allowance not stated

No

No

Copilot Pro

$10

1,500

No

Yes

Copilot Pro+

$39

7,000

Yes

Yes

Copilot Max

$100

20,000

Yes

Yes

Copilot Business

$19 per seat

1,900 per user, pooled

Yes

Yes

Copilot Enterprise

$39 per seat

3,900 per user, pooled

Yes

Yes


........


Plan prices and credit allowances are those listed in GitHub's documentation on November 19, 2026. Code completions and next edit suggestions do not consume AI credits, so the allowance is spent on chat, agent and coding-agent requests, which is where a model such as Neon would be used.


··········


GEMINI 4 NEON PRICE PER 1M TOKENS COMPARED WITH OTHER COPILOT MODELS.


GitHub's announcement does not print Neon's per-token rates, and the GitHub models and pricing reference had not yet added a Neon row when it was checked on November 19, 2026. Since the changelog states that the model is billed at provider list pricing, the relevant figures are Google's. The table below uses Google's global standard rates for the Gemini 4 models, as listed on the official Google Cloud pricing page on November 19, 2026, and GitHub's own published Copilot rates for the Claude and OpenAI models.


........


Model in Copilot

Input

Cached input

Output

Source of rate

Gemini 4 Neon

$3.00

$0.15

$15.00

Google list price

Gemini 4 Hydrogen

$1.50

$0.075

$7.50

Google list price

Gemini 4 Carbon, to Nov 30

$2.00

Not listed

$10.00

Google list price

Gemini 4 Carbon, from Dec 1

$4.00

Not listed

$20.00

Google list price

Claude Opus 5.5

$4.00

$0.20

$20.00

GitHub pricing table

Claude Sonnet 5.5

$2.00

$0.10

$10.00

GitHub pricing table

GPT-6.1 Terra, up to 272K

$1.00

$0.05

$5.00

GitHub pricing table


........


All prices are in USD per million tokens. Google's page lists a cache write rate for Neon of $3.75 per million tokens, effective January 12, 2027, and GitHub's table lists cache write rates of $5.00 for Claude Opus 5.5 and $2.50 for Claude Sonnet 5.5. Google states that Neon's pricing is flat across its full 1,048,576-token context window, while GPT-6.1 Terra moves to a higher tier above 272,000 input tokens in Copilot.


The comparison with Gemini 4 Carbon depends on the date. Until November 30, Carbon's promotional rates of $2 and $10 are a third below Neon's. From December 1, Carbon returns to $4 and $20, and Neon becomes 25% cheaper per token on both input and output. Google also claims that Neon costs 32% less per task than Carbon at its default high thinking level, because it completes comparable work with fewer tokens; that is a benchmark-based claim from Google and the saving on a given codebase will vary.


··········


WHAT A GEMINI 4 NEON CODING SESSION COSTS IN COPILOT AI CREDITS.


To show what the list prices mean for a subscriber, the example below models one agent session in Copilot with 40 model turns. Each turn sends 60,000 input tokens, of which 80% are served from the cache because the repository context and earlier conversation repeat, and returns 2,500 output tokens including thinking. Over the session that adds up to 2.4 million input tokens (480,000 fresh and 1.92 million cached) and 100,000 output tokens. Cache write charges are left out for every model, since Neon's cache write rate only applies from January 12, 2027, and the same token counts are applied to each model, which ignores differences in how many tokens each model actually needs.


........


Model

Fresh input

Cached input

Output

Session cost

AI credits

Gemini 4 Neon

$1.44

$0.29

$1.50

$3.23

323

Gemini 4 Hydrogen

$0.72

$0.14

$0.75

$1.61

161

Claude Opus 5.5

$1.92

$0.38

$2.00

$4.30

430

Claude Sonnet 5.5

$0.96

$0.19

$1.00

$2.15

215

GPT-6.1 Terra

$0.48

$0.10

$0.50

$1.08

108


........


At about 323 credits per session of this size, a Copilot Pro+ allowance of 7,000 credits covers roughly 21 Neon sessions a month and a Copilot Max allowance of 20,000 credits covers about 61. On Copilot Business, the 1,900 credits per user correspond to about five such sessions, and on Copilot Enterprise the 3,900 credits per user correspond to about 12, although business allowances are pooled across the organization, so heavy and light users offset each other. Usage beyond the allowance is billed at the same per-token rates.


The arithmetic places Neon between the two Claude models in Copilot: it costs about half again as much as Claude Sonnet 5.5 for the same token volume and about a quarter less than Claude Opus 5.5. Against Gemini 4 Hydrogen, Neon costs exactly twice as much per token, so developers who were happy with Hydrogen for routine agent work have a direct cost reason to keep it as the default and switch to Neon for harder tasks.


··········


WHAT CHANGES FOR GITHUB COPILOT USERS AND ADMINISTRATORS.


For individual developers on Copilot Pro+ or Max, Neon is a new option in the model picker as soon as the rollout reaches the account, with no setting to change. Developers on Copilot Pro who want Google's newest model inside Copilot need to move to Pro+, which raises the monthly price from $10 to $39 and the allowance from 1,500 to 7,000 credits; our comparison of Copilot Pro and Pro+ sets out the other differences between the two plans. Pro users keep access to Gemini 4 Hydrogen, which GitHub added at $1.50 and $7.50 per million tokens the week before.


For organizations, the practical step is a policy review. Under default enablement Neon switches on by itself, and at list prices one long agent session can use a sixth of a Business user's monthly credits, so teams with tight AI credit budgets may want to restrict it to specific users or keep Hydrogen and Sonnet-class models as the defaults. Our guide to AI coding agent pricing across Copilot, Claude Code, Codex and Cursor covers how Copilot's plan prices compare with the other coding tools.


The December 1 change to Carbon's pricing also affects model choice inside Copilot. Teams that adopted Carbon during its promotional period will see its per-token cost double at the start of December, at which point Neon offers lower list prices on both input and output. Google's launch details, benchmark figures and API terms for Neon are covered in our Gemini 4 Neon launch article.


··········


GEMINI 4 NEON IN COPILOT: CONFIRMED FACTS AND OPEN QUESTIONS.


GitHub has confirmed general availability, the four eligible plans, the full list of clients, billing at provider list pricing and the default enablement behavior for Business and Enterprise. Google has published Neon's list prices of $3, $0.15 and $15 per million tokens, a flat rate across the 1M-token window, and a cache write rate that starts in January 2027.


Several points are still open. GitHub's pricing reference did not yet show a Neon row on November 19, so the exact cached-input and cache-write treatment inside Copilot has not been documented by GitHub itself. GitHub gave no reason for leaving Copilot Pro out, and did not say whether that could change. Google's per-task savings against Carbon and competing models come from Google's selection of benchmarks, and Copilot users will only see the real cost difference in their own credit usage reports. For Pro+ and Max subscribers the model is available now at a price that sits between Claude Sonnet 5.5 and Claude Opus 5.5, and from December 1 it undercuts Gemini 4 Carbon per token.


··········


FOLLOW US FOR MORE.


·····


DATA STUDIOS


·····


datastudios.org


bottom of page