Google: Gemini 4 Argon

Google's most capable Gemini model, built to reason longer on hard coding, agentic, math and science problems

gemini-4-argon

AvailableGoogleChatCached input 95% lessNew

Input / Output-50%: $2.00 / $10.00$1.00 / $5.00
Context: 1M
Released: Oct 13, 2026
Max output: 64K
Modalities: VisionChat
Capabilities: Tool usePrompt CacheReasoning

About Gemini 4 Argon

Gemini 4 Argon is Google's most capable Gemini model, released in October 2026 for complex coding, agentic, math and science work and for real-world software engineering and knowledge tasks. Google built it to reason longer on hard problems: input, thinking and output share one context window, so it can spend more of that window on thinking than earlier Gemini models. It reads text, images, video, audio and PDF.

Where it works well

Built to reason longer on hard problems, with low, medium and high thinking levels to trade depth against latency.
Aimed at real-world software engineering and multi-step agent work as well as math and science problems.
Reads text, images, video, audio and PDF in one request and answers in text.
Calls functions and returns schema-bound JSON for structured agent steps.

When to choose another model

Flash models answer faster and cost less for simple, high-volume tasks that do not need deep reasoning.
Explicit context caching is not supported; repeated long prompts rely on implicit caching only.
Temperature, top_p and top_k are ignored, so sampling cannot be tuned for more varied output.

Getting started

Create API key
Create a key in Console, then use it with every model on the platform.

Send your first request

Copy the example for your language and run it against the endpoint.

Text to textGemini API

POST/v1beta/models/gemini-4-argon:generateContent

API format:

curl "https://api.tokenlab.sh/v1beta/models/gemini-4-argon:generateContent" \
  -H "Content-Type: application/json" \
  -H "x-goog-api-key: sk-xxx" \
  -d '{
    "contents": [
      {"role": "user", "parts": [{"text": "Hello!"}]}
    ]
  }'

Pricing

The Verified price applies to Verified, which costs less on most models. The Official price is the model maker's published price and applies to the more reliable Official route. Auto bills the route that completes the request.

Default spec

per 1M tokens

Official price: Input $2.00 / Output $10.00 / Cache read $0.10
Verified price: Input $1.00 / Output $5.00 / Cache read $0.05
Discount: -50%

Prompt cache pricing

Cache read

Official price: $0.10
Verified price: $0.05
Discount: -50%

Tool fees

Charged per call, only when the model uses the tool.

Web search

Official price: $0.014/search
Verified price: $0.007/search
Discount: -50%

	Official priceper 1M tokens	Verified priceper 1M tokens	Discount
Default spec	Input $2.00 / Output $10.00 / Cache read $0.10	Input $1.00 / Output $5.00 / Cache read $0.05	-50%
Prompt cache pricing
Cache read	$0.10	$0.05	-50%
Tool feesCharged per call, only when the model uses the tool.
Web search	$0.014/search	$0.007/search	-50%

Usage & activity

Success rate is the share of requests that completed. Latency is how long a full response takes; P95 means 95% of requests finished within that time.

Usage & availability

Last 24 hours

Requests
Success rate
P95 latency
Total tokens

Model performance

Last updated Oct 14, 2026, 6:33 PM UTC

Metrics appear once privacy and data volume thresholds are met.

Data is based on aggregate user requests, excluding status checks.

Open in Console

Open Gemini 4 Argon in Console with a prompt ready to edit or send.

Help me try gemini-4-argon with a short message at /v1beta/models/gemini-4-argon:generateContent. Show the reply, latency, and cost.

Use cases

Best for

Reasoning
Vision

Hard coding tasks
Give it a failing build, a migration or a refactor that spans many files and let it plan the change before it writes code.
Agents with many steps
Drive an agent that calls tools, reads the results and decides the next step across a long task.
Math and science problems
Work through proofs, derivations and quantitative analysis at the high thinking level when accuracy matters more than speed.
Multimodal analysis
Combine documents, screenshots, recordings and video in one request and ask for a written analysis or structured result.

Prompt examples

Here is the stack trace and the three modules it touches. Find the root cause, propose a fix, and list the tests that should change.

Prove that the sum of the first n odd numbers is n squared, then generalize the argument to arithmetic sequences.

Watch this product demo video and the attached spec PDF, then list every place where the demo differs from the spec.

This model has conditional pricing. Monthly token totals alone cannot produce a reliable estimate; use the detailed pricing for the request specification, cache, and applicable time window.

FAQ

What is Gemini 4 Argon?

Google's latest and most capable Gemini model, released in October 2026. Google describes it as built to reason longer on hard problems in coding, agentic, math and science work, and for real-world software engineering and knowledge tasks.

How do I control how much it thinks?

Set a thinking level: low uses the fewest tokens and answers fastest, medium is the default and Google's recommendation for most tasks, and high spends the most tokens for complex reasoning, hard math and science. In Chat Completions, use reasoning_effort with the same three values.

Does it support temperature and top_p?

No. Google states that temperature, top_p and top_k are ignored for this model, so requests that set them still succeed but those values have no effect. Remove them from existing Gemini integrations to avoid confusion.

Should I use it instead of Gemini 3.8 Flash?

Use Argon where a task rewards longer reasoning, such as hard bugs, multi-step agents or math and science problems. For fast, high-volume work such as chat, extraction or routine coding, Gemini 3.8 Flash is quicker and cheaper; compare both on your own evaluation.

How much does Gemini 4 Argon cost?

On TokenLab, Gemini 4 Argon costs Input $1.00 / Output $5.00 per 1M tokens. The pricing table above shows the full breakdown. Rates depend on the billing unit, specification, and usage. Compare matching conditions in the model's detailed pricing; a single rate does not determine the total cost.

What are the context window and output limit of Gemini 4 Argon?

Gemini 4 Argon accepts up to 1,048,576 tokens of context and returns up to 65,536 tokens in one response.

Which endpoint should Gemini 4 Argon use?

Use https://api.tokenlab.sh/v1beta/models/gemini-4-argon:generateContent for Gemini 4 Argon. The request example below shows the matching code shape.

Which operations does Gemini 4 Argon support?

Gemini 4 Argon supports Text to text. Select an operation above to see its endpoint and request example.