GPT-Image 2.6 Flare & Sunburst are liveImages from $0.025 each

Gemini 4 Neon API

NewLLM

The Gemini 4 Neon API is Google's middle Gemini 4 route, between Hydrogen for fast everyday execution and Carbon for the hardest long-horizon projects, with a 1M context, thinking levels up to xhigh, and text, image, video, audio and PDF input.

What are we making today?

Bring an idea, a question, or a first draft.

Enter to send · Shift + Enter for a new line

01

What this model is for

The Gemini 4 Neon API is Google's middle Gemini 4 route, between Hydrogen for fast everyday execution and Carbon for the hardest long-horizon projects, with a 1M context, thinking levels up to xhigh, and text, image, video, audio and PDF input.

Gemini 4 Neon is the Gemini 4 route Google places between Hydrogen and Carbon, and the model card says it is adapted from Carbon, so the architecture, training data and safety work are shared. OmniaKey lists it on the same flat Gemini 4 structure as its siblings, quoted at the Google Pro ratio, over a 1M-token window that accepts text, images, video, audio and PDF files.

What separates Neon from Carbon is the thinking budget: Neon defaults to the high level and accepts xhigh, while Carbon stops at high. That makes Neon the route for work that needs sustained reasoning — a long refactor review, a document set with a genuine question, a multi-turn plan — where the failure path is understood but the job runs long.

Use Neon when a task needs more depth than Hydrogen without stepping up to Carbon; keep Hydrogen for high-volume classification and Carbon for multi-service debugging where the failing path is still unknown.

Best-fit coding workloads

Sustained reasoning over a document set

Send the contracts, specs or papers together with the question, and ask for the answer plus the passage that supports each claim. Practical tip: state which document is authoritative, because the model quotes what it is given and will not resolve a conflict the prompt never names.

Long refactors with the plan attached

Give it the module, its callers and the target contract, then ask for a staged plan with the test that proves each stage. Practical tip: request the second stage only after the first one passes, so a long horizon does not collapse into one unverifiable change.

Multi-turn technical planning

Use it for design conversations that run across many turns — schema changes, migrations, incident postmortems — where context keeps accumulating. Practical tip: keep a short running summary in the latest message, since a large window still degrades when its middle is never revisited.

02

Technical specification

Catalog facts and integration details verified for this route.

Model ID
gemini-4-neon
Context
1M
Inputs and tools
Vision · Tool use
API endpoint
/v1/chat/completions
Cache hit / 1M
$0.06 USD
Coding clients
Cursor · Cline · Aider

Limits to plan around

  • Neon sits above Carbon on our rate card, so it is a poor default for routine traffic: reach for it when the work is long, not when it is merely unfamiliar.
  • Explicit context caching is not available on this route. Only the implicit cache discounts a repeated prefix, and that needs a prompt of at least four thousand tokens to engage.
  • We have not yet run our own verification suite on Neon, so treat the sibling Gemini 4 pages as the reference until Neon's own checks are published.

03

Compare nearby models

Compare workload fit and current listed token rates before choosing a route.

ModelBest forContextInputOutput
Gemini 4 NeonThis page
gemini-4-neon
Google's middle Gemini 4 route, adapted from Carbon, with thinking levels up to xhigh and multimodal input.1M
$1.2 USD
Official price $3 USD
$6 USD
Official price $15 USD
Gemini 4 Carbon
gemini-4-carbon
The hardest long-horizon Gemini 4 route; its thinking level stops at high and it is listed below Neon today.1M
$0.8 USD
Official price $2 USD
$4 USD
Official price $10 USD
Gemini 4 Hydrogen
gemini-4-hydrogen
The fast everyday Gemini 4 route for high-volume loops; pick it when depth matters less than latency.1M
$0.8 USD
Official price $2 USD
$4 USD
Official price $10 USD

04

SDK quickstarts

Use one base URL and model ID from cURL, Python, or TypeScript.

Streaming chat completion
curl https://api.omniakey.com/v1/chat/completions \
  -H "Authorization: Bearer $OMNIAKEY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gemini-4-neon",
    "messages": [
      {"role": "user", "content": "Review this pull request"}
    ],
    "stream": true
  }'

05

Estimate token cost

Enter expected input and output tokens to compare current listed rates.

OmniaKey estimate
$0.24
Official-rate estimate
$0.60
Estimated savings
$0.36

Estimate covers input and output tokens only. Cache usage, retries, tools, and client-side token accounting can change the final charge.

07

Sources and verification

Model facts and links were checked on 2026-11-19. Provider documentation remains the authority for upstream behavior.

Frequently asked questions

How much does Gemini 4 Neon cost on OmniaKey?

OmniaKey charges $1.2 USD per 1M input tokens and $6 USD per 1M output tokens, in USD. Official prices are $3 USD and $15 USD; you save 60%.

What is the context window for Gemini 4 Neon?

The listed context window is 1M. Actual usable limits can also depend on the client and upstream request format.

Where does Gemini 4 Neon sit in the Gemini 4 family?

Between Hydrogen for fast everyday execution and Carbon for the hardest long-horizon projects. Google's model card says Neon is adapted from Carbon, so it inherits that model's architecture and training work, while its default thinking level and available levels differ.

Is Neon cheaper than Carbon?

Not today: Neon is listed above Carbon, partly because Carbon is still on its introductory rate. Google's standard Carbon rate is scheduled to rise in December, which reverses the order, so re-check both pages before you commit a budget.

What can I send to Gemini 4 Neon?

Text, images, video, audio and PDF files across a 1M-token context, with text output, tool calling and thinking levels from low to xhigh. Output billing includes the thinking tokens.

Does Neon support explicit context caching?

No. Only implicit caching applies, and it discounts a repeated prefix of at least four thousand tokens when it hits. Prompts with a stable opening still benefit; cache objects you manage yourself are not available.