Final Router

Google · model

Gemini 4 Argon

Reasoning, Huge context, Multimodal. Called through Final Router's one OpenAI-compatible endpoint at Google's list price, with no markup on tokens.

Model id: google/gemini-4-argon
Input price$2per 1M tokens
Output price$10per 1M tokens
Context window1.048576Mtokens
Typical speed1.5smedian
Quality score97/100composite
ProviderGoogleprocessed in US
Cached input$0.1per 1M, cache read
Cache writeIncludedautomatic, no write fee

Note: From 2026-12-01 the provider's price rises to $4.00 in and $20.00 out, per 1M, and we bill the new price from that day.

Verified 2026-11-02Google's own list price, with no markup on tokens - a 5.5% fee applies only when you add credits.

Call Gemini 4 Argon in one line

Gemini 4 Argon speaks through the same OpenAI-compatible endpoint as every other model. Point your existing client at Final Router, set the model id, and you also get automatic fallback, spend caps and usage tracking - none of which the provider gives you directly.

Call Gemini 4 Argon
from openai import OpenAI

client = OpenAI(
    api_key="fr_live_...",                     # your Final Router key
    base_url="https://finalrouter.com/api/v1"  # the only line you change
)

resp = client.chat.completions.create(
    model="google/gemini-4-argon",
    messages=[{"role": "user", "content": "Hello!"}],
)
print(resp.choices[0].message.content)

Data & residency

The Gemini API endpoint we call is global, which in practice means processing outside the EU. Vertex AI offers European regions; that is a different integration and not this one.

Training: not used to train the model. Retention: Paid tier is not used for training; the free tier is. Check which key you connected.

Google's policy, read 2026-08-21 →

Route it, don't pin it

Name Gemini 4 Argon as your first choice and let a policy fall back to another model when the provider is down - your request still gets answered, and the log shows exactly what happened.

Start free

Gemini 4 Argon FAQ

How much does Gemini 4 Argon cost through Final Router?
$2 per million input tokens and $10 per million output tokens - the provider's own list price, with no markup on tokens. A 5.5% fee applies only when you add credits.
What is Gemini 4 Argon's context window?
1,048,576 tokens.
Does Gemini 4 Argon train on my data?
No. On the endpoint Final Router calls, Google does not train on your prompts. Data is processed in US.
How do I call Gemini 4 Argon?
Point any OpenAI-compatible client at https://finalrouter.com/api/v1 with your Final Router key and set model to "google/gemini-4-argon". Your existing code does not change.

Compare with

All comparisons

gemini-4-argonLatency and quality are seed figures used for routing order, not billing. Prices are what you are charged.