Gemini 4 Hydrogen API

NewLLM

The Gemini 4 Hydrogen API is the newest route in Google's Gemini 4 family on OmniaKey, billed on the same rate card as Carbon and Argon: 1M context, image input and tool use, with one flat band for the whole window.

Read the docs Get your API key

Input · USD / 1M tokens$0.8 USDOfficial price~~$2 USD~~
Output · USD / 1M tokens$4 USDOfficial price~~$10 USD~~
Savings60%
Context1M
View tool pricing and billing rules

What are we making today?

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

What this model is for

Gemini 4 Hydrogen is the third route in Google's Gemini 4 family on OmniaKey, after Carbon and Argon, and the newest of the three. Google bills the trio on one rate card, and the gateway prices Hydrogen at Google's Pro ratio, so it lands on the same input, output and cached-input rates as its siblings. The catalog lists a 1M-token context window, image input and tool use, with one flat band across the whole window.

Because the price is identical, the reason to pick Hydrogen is the snapshot, not the bill. In our 12 November 2026 runs a merge-intervals coding prompt returned runnable Python on all three calls with a median of 206 output tokens, a JSON-mode request came back as a parseable object, tool calling produced a valid get_weather call, and image input described a supplied cover correctly. Those checks show the route is ready for structured and multimodal work; they also mean switching from Carbon or Argon changes the answers, not the cost.

Move long-context multimodal work here when you want the newest Gemini 4 snapshot at the Carbon and Argon rate; keep Gemini 3.8 Flash for high-volume loops where latency and the bill lead.

Best-fit coding workloads

Drop-in Gemini 4 upgrade

Change the model id in an existing Carbon or Argon integration and replay its evaluation set before touching any prompt. Because the rate card matches, the only variable is whether the newer snapshot answers your hardest cases better.

Long document review

Load a specification, a contract or a set of logs into the 1M window and ask which clauses conflict, by page. Practical tip: require page numbers next to every claim so a reviewer can check the source line before signing.

Structured extraction with tools

Turn tickets or emails into fixed-key JSON and let the model call a lookup tool for missing fields. Practical tip: pin the date format, state that the newest message wins, and validate the object before a production batch runs.

Hands-on notes

Checked on 2026-11-12 through the OmniaKey API with one standard coding prompt (median of 3 runs) and three capability probes.

Output tokens for our standard coding prompt, reasoning included: 206
Cost of that output at current prices: $0.00082
JSON mode: Passed
Tool calling: Passed
Image input: Passed

On 12 November 2026 we ran a merge-intervals coding prompt three times: every call returned runnable Python with the requested edge cases, and the median reply was 206 output tokens.
The same session passed a JSON-mode request, produced a valid get_weather tool call with the right city argument, and correctly described a supplied cover image.

Technical specification

Catalog facts and integration details verified for this route.

Model ID: gemini-4-hydrogen
Context: 1M
Inputs and tools: Vision · Tool use
API endpoint: /v1/chat/completions
Cache hit / 1M: $0.04 USD
Coding clients: Cursor · Cline · Aider

Limits to plan around

The rate card is identical to Carbon and Argon, so switching cannot lower a bill; the only reason to move is wanting the newer snapshot.
Our checks cover text, JSON, tool calling and image input; we have not run long agent traces or audio input on this snapshot, so pilot those workloads before moving them over.
Nothing is aliased: clients must ask for the new id explicitly, and a fleet that mixes the three Gemini 4 routes makes per-snapshot cost attribution harder.

Compare nearby models

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

Model	Best for	Context	Input	Output
Gemini 4 HydrogenThis page gemini-4-hydrogen	Newest Gemini 4 snapshot on the shared Carbon and Argon rate card; the drop-in upgrade for existing Gemini 4 deployments.	1M	$0.8 USD Official price $2 USD	$4 USD Official price $10 USD
Gemini 4 Argon gemini-4-argon	Same window and same rates as Hydrogen; stay here unless you specifically want the newer Gemini 4 snapshot.	1M	$0.8 USD Official price $2 USD	$4 USD Official price $10 USD
Gemini 3.8 Flash gemini-3.8-flash	The cheaper Flash route for high-volume loops; move up to a Gemini 4 model only when a task needs the long window or harder reasoning.	1M	$0.45 USD Official price $0.75 USD	$2.25 USD Official price $3.75 USD

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-hydrogen",
    "messages": [
      {"role": "user", "content": "Review this pull request"}
    ],
    "stream": true
  }'

Estimate token cost

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

Input tokensOutput tokens

OmniaKey estimate: $0.16
Official-rate estimate: $0.40
Estimated savings: $0.24

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

Related coding guides

Deeper selection, setup, and troubleshooting guidance for this model family.

Sources and verification

Model facts and links were checked on 2026-11-12. Provider documentation remains the authority for upstream behavior.

Frequently asked questions

How much does Gemini 4 Hydrogen cost on OmniaKey?

OmniaKey charges $0.8 USD per 1M input tokens and $4 USD per 1M output tokens, in USD. Official prices are $2 USD and $10 USD; you save 60%.

What is the context window for Gemini 4 Hydrogen?

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

How does Gemini 4 Hydrogen compare in price with Carbon and Argon?

It matches them. Google bills the three Gemini 4 routes on one rate card, and OmniaKey publishes the same input, output and cached-input rates for Hydrogen as for Carbon and Argon, so the choice is about which snapshot you want rather than what it costs.

Should we migrate an existing Gemini 4 Argon integration?

Only if you want the newer snapshot. The request format and the rate card are unchanged, so the work is a model id swap plus one replay of your evaluation set; if Argon already passes your cases, there is no financial reason to move.

What input types can Gemini 4 Hydrogen handle?

Text and images, across a 1M-token window, with tool calling, JSON mode and streaming. Our own checks covered text, JSON mode, a tool call and one image, and every one of them behaved as documented.

How was Gemini 4 Hydrogen verified on OmniaKey?

We ran a merge-intervals coding prompt three times on 12 November 2026 and recorded the median output length, then tested JSON mode, tool calling and image input on the same route; the results are the table above and the notes in this guide.