Gemini 4 Neon Is GA at $3/$15, and Lands in GitHub Copilot Same Day
Google released Gemini 4 Neon (gemini-4-neon) as generally available on 18 November 2026, per the Gemini API release notes. It is a paid-tier model aimed at "software engineering, knowledge work, and science workflows", with a 1M-token context window and four thinking levels. GitHub put it in Copilot the same day. It is the fourth Gemini 4 model to reach GA after Argon, Carbon and Hydrogen, and it slots in between them on price.
What happened
From Google's pricing page and model page:
| Standard | Batch / Flex | |
|---|---|---|
| Input (per 1M) | $3.00 | $1.50 |
| Cached input (implicit only) | $0.15 | $0.075 |
| Output, incl. thinking tokens (per 1M) | $15.00 | $7.50 |
| Context window | 1,000,000 | 1,000,000 |
| Long decode | up to 1M tokens | up to 1M tokens |
A few details matter for billing. Neon is not on the free tier at all, and Google's page is explicit that every token is counted, "including thinking tokens and internal tool system instructions and function declarations when using grounding". It supports implicit caching only (no explicit cache), and offers four thinking levels: low, medium, default high, and xhigh. Because output billing includes thinking tokens, the default high level is where an unwatched bill grows.
Where it sits against its siblings on the pricing tracker:
| Model | Input | Output | Notes |
|---|---|---|---|
gemini-4-carbon |
$2.00 | $10.00 | Doubles to $4/$20 on 1 December |
gemini-4-neon |
$3.00 | $15.00 | New, no free tier |
gemini-4-hydrogen |
$1.50 | $7.50 | Free tier available |
So Neon is priced above today's Carbon but below where Carbon lands on 1 December, when Argon and Carbon double. After that date, Neon at $3/$15 becomes the cheaper of the two frontier Gemini options. Hydrogen stays the budget tier.
In GitHub Copilot, Neon is generally available to Pro+, Max, Business and Enterprise users across VS Code, Visual Studio, the Copilot CLI, the coding agent, github.com, GitHub Mobile, JetBrains, Xcode and Eclipse. GitHub bills it at "provider list pricing under usage-based billing", and Google reports it responds faster and makes fewer unrequested edits than Carbon on coding tasks. As with the other recent Copilot model additions, it is switched on automatically unless an admin has turned off the default.
What changes for engineers
- The default thinking level is the cost lever. Output tokens include thinking tokens, and the default is
high. For routine agent calls,lowormediumcan cut the output bill substantially; reservexhighfor the hard tasks that justify it. - No free tier means no free evals. Unlike Hydrogen, you cannot try Neon without a billing account. Budget for the eval run itself.
- Batch and Flex halve the token price. Non-interactive work (nightly summaries, bulk extraction, eval backlogs) runs at $1.50/$7.50. That is cheaper than Carbon's Standard rate today and well under Carbon's December price.
- Caching is implicit only. If your workload relied on explicit cache control on other models to pin a long system prompt, that lever is not here; you get implicit caching at $0.15 per 1M and no guarantee of a hit.
What I ran
Nothing. I have no Gemini API key on this machine and Neon has no free tier, so there are no latency or quality numbers here. The prices and limits above are from Google's own pricing and model pages, fetched on 19 November 2026. When I run a fixed task set across the Gemini 4 line I will publish the results separately. I have updated the model release tracker and the pricing tracker with the verified numbers.
What to do this week
- If you use Copilot, decide before it defaults on. Enterprise and Business admins who want to vet models first should check the model policy in Copilot settings; new models enable automatically otherwise.
- Set an explicit thinking level in API calls. Do not leave it at the
highdefault for cheap, high-volume calls. - Move eligible jobs to Batch or Flex for the half-price token rate, and compare against Hydrogen before assuming you need a frontier model at all.
FAQ
How much does Gemini 4 Neon cost?
On the Standard tier, $3.00 per million input tokens and $15.00 per million output tokens, with implicit cached input at $0.15. Batch and Flex are half that: $1.50 input and $7.50 output. Output billing includes thinking tokens. There is no free tier.
What is the context window of Gemini 4 Neon?
A unified 1,000,000-token context window, with long decode (output generation) of up to 1M tokens, per Google's model page.
How does Gemini 4 Neon compare to Carbon and Hydrogen?
Neon ($3/$15) is priced above Carbon's current $2/$10 and above Hydrogen's $1.50/$7.50. On 1 December 2026, Carbon doubles to $4/$20, which makes Neon the cheaper frontier Gemini option after that date. Hydrogen remains the budget tier and keeps a free tier that Neon does not have.
Related: AI Model Release Tracker, LLM API Pricing Tracker and Gemini 4 Argon and Carbon API Prices Double on 1 December.
Spot an error or something out of date? Email me and I will fix it and log it on the corrections page.
I'm Ajay Kumar — I build and operate PandaStack, an open-source Firecracker microVM cloud for AI agents, and write about the infrastructure and tooling around it.
Need this kind of infrastructure work? See what I do or email hello@ajayk.sh.
Related
OpenAI GPT-6.2 Astra and Sol: Prices, Limits, Background Windows
OpenAI released gpt-6.2-astra and gpt-6.2-sol on 17 November 2026. Astra is 20% cheaper than gpt-6-astra, Sol keeps 6.1 Sol's price, and a new background tier cuts token rates 30 to 40%. What changed, what it costs, and what to test.
5 minNov 13, 2026NEAR AI Breached Through an SGLang RCE: No CVE, No Upstream Fix Yet
NEAR AI says an attacker used an unpatched remote code execution flaw in SGLang on 8 November 2026 to reach ~100 users' account data, 18 legacy agents and requests running in its TEE. No CVE or SGLang advisory exists. What SGLang operators should do now.
6 minNov 10, 2026Alphabet Q3 2026: $51.4B Capex, Cloud Up 103%, Free Cash Flow at -$9.1B
Alphabet's Q3 2026 results: $51.4 billion of quarterly capex, Google Cloud revenue up 103% to $30.7 billion, negative free cash flow of $9.1 billion, an $802 billion commitments book and new debt and equity to pay for it.
7 min