Gemini 4 Argon API in India

In Indian rupees, the currency your card is charged in, with the published US dollar rate beside each.

For scale: 1,000 requests of about 3,000 tokens in and 1,000 out cost ₹1,615.04 — a long prompt and a page of answer each. Every rate is published and stays the same from one request to the next. See every model.

Prices are our published US dollar rates converted at ₹98.00 to the dollar, our rate for October — the same rate your card is charged.

Argon is the first Gemini 4 model, and Google aims it at long, multi-step work in finance, law, coding and science. Independent tests mostly agree on where it is strong: it ranked first of 43 models on the Vals Index, which covers finance, legal, tax and coding tasks, and tied for third on the Artificial Analysis Intelligence Index, level with GPT-6 Astra and Claude Fable 5.1.

Its most useful trait may be restraint. On Artificial Analysis’s AA-Omniscience test, which penalises wrong answers but not refusals, it hallucinated 15% of the time — the lowest of any model scoring that high on the index. It is more willing than its rivals to say it does not know.

It reads text, images and PDFs in a window of about a million tokens, and you choose how hard it thinks, from low to high. Web search, reading a URL and running Python are built in: ask for one as a tool, and sources come back as citations.

• Finance, legal and tax analysis across long documents.
• Research questions where a confident wrong answer is worse than none.
• Grounded answers with sources, using its built-in web search.
• Agents that call your tools over many steps.

• Not the strongest coder. Google’s own table puts it behind Claude Opus 5.5 on Terminal-Bench 4.0, and independent reviews say it trails Anthropic’s newest models on agentic coding.
• It refuses more readily than most, so an application should have a plan for questions it will not answer.
• Google’s launch pricing ends on 30 November, and the rate on this page goes up with it.
• Video input and Google’s million-token answers are not available through our API yet; answers are capped at about 65K tokens.

Roar AI provides access to Gemini 4 Argon through the Roar AI gateway, charged in Indian rupees on one bill with every other model. Requests from India go through our gateway in Bangalore. Developers can use Gemini 4 Argon alongside other supported models without a separate integration for every provider.

Use the OpenAI SDK you already have — point it at Roar AI and name the model. Read the docs.

from openai import OpenAI

client = OpenAI(base_url="https://api.roar-ai.com/v1", api_key="roar_live_…")

reply = client.chat.completions.create(
    model="gemini-4-argon",
    messages=[{"role": "user", "content": "Summarise this contract in plain English."}],
)

Gemini 4 Argon is a frontier model, and requests are processed on international infrastructure, outside India. The model directory says where each model is processed.

Claude Sonnet 5.5Similar priceGPT-6.1 SolSimilar priceGPT-6 SolSimilar price

Every Gemini model: Gemini API. Explore all AI models available in India