Gemini 4 Carbon API

Gemini 4 Carbon is a model from Google's Gemini 4 series, built for coding, agents and complex reasoning. It has one shared 1,048,576-token window for input, thinking and output. Thinking depth has three levels (low, medium - the default, high). It supports tool calls and JSON schema output. It answers with text. In Jantaris you call it with the same OpenAI SDK as every other model, without a separate Google account.

A single answer is capped at 32,768 tokens without streaming and 65,536 with streaming, and thinking counts toward that cap. For long answers set stream: true.

The same series has two cheaper models: Gemini 4 Hydrogen for fast, high-volume work and Gemini 4 Argon. Pick Carbon where code quality and long multi-step tasks matter.

Best for

• Writing and fixing code in large repositories
• Tool-using agents with many steps
• Analysis of very long documents and data extraction to JSON
• Head-to-head tests with Claude Opus 5.5 and GPT-6.1 Sol on your own tasks

How to call it

from openai import OpenAI

client = OpenAI(base_url="https://jantaris.com/v1", api_key="sk-jt-...")
resp = client.chat.completions.create(
    model="gemini-4-carbon",
    stream=True,
    reasoning_effort="high",
    messages=[{"role": "user", "content": "Find the bug in this function and suggest a fix."}],
)
for chunk in resp:
    print(chunk.choices[0].delta.content or "", end="")

FAQ

What is the Gemini 4 Carbon API price?

$2.16 / $10.80 per 1M tokens. You pay from one balance, only for the tokens you use.

How is it different from Gemini 4 Hydrogen and Argon?

Carbon targets coding, agents and hard reasoning and has the highest token price of the three. Gemini 4 Hydrogen and Gemini 4 Argon have the same window size and fit better where a low price on large volumes of text matters.

How long can an answer be?

Up to 32,768 tokens without streaming and 65,536 with streaming, thinking included.

Does Carbon generate images or audio?

No, it answers with text only.

How to call Gemini 4 Carbon via API

from openai import OpenAI

client = OpenAI(base_url="https://jantaris.com/v1", api_key="sk-jt-...")
resp = client.chat.completions.create(
    model="gemini-4-carbon",
    messages=[{"role": "user", "content": "Write a haiku about Baltic amber."}],
)
print(resp.choices[0].message.content)

curl https://jantaris.com/v1/chat/completions \
  -H "Authorization: Bearer $JANTARIS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "gemini-4-carbon", "messages": [{"role": "user", "content": "Write a haiku about Baltic amber."}]}'

import OpenAI from "openai";

const client = new OpenAI({ baseURL: "https://jantaris.com/v1", apiKey: process.env.JANTARIS_API_KEY });
const resp = await client.chat.completions.create({
  model: "gemini-4-carbon",
  messages: [{ role: "user", content: "Write a haiku about Baltic amber." }],
});
console.log(resp.choices[0].message.content);