← All models
TextNew

Gemini 4 Argon

Google's newest Gemini 4 model for hard tasks: plans, code review, step-by-step reasoning and long texts. You can attach pictures.

Good for: hard questions, code, analysis, long texts

Smart assistantUp to 10 of your photos at once
Start creating $2 / 1M input · $10 / 1M outputpay per token

A token is about 4 characters, an attached picture is 560 tokens. A reserve for your question and up to 16384 answer tokens is held at start; after the answer only the actual tokens (an estimate) are charged and the rest returns to your balance at once.

For developers →

Your question

1
Write your question
2
Add your photooptional

The AI builds on it. You can skip this.

Drop a photo here or clickTap to choose a photo JPG, PNG or WebP, up to 30 MB
3
Thinking leveloptional

You can leave it as is — the defaults work well.

Thinking level ?
Sign in to create

usually under a minute

How to get a better result

Be specific

Instead of "write a text" say "write a Telegram post about walking, 3 paragraphs".

Say who it is for

"Explain how a neural network works to a schoolchild".

Not right? Refine

Say what to change and press "Create" again.