Felo API PlatformFelo API Platform
Back to models
Googlemessages

Gemini 4 Carbon (batch) API

Gemini 4 Carbon is Google's multimodal frontier model for long-horizon agentic work, with 1 million token context and adjustable reasoning effort.

Gemini 4 Carbon is a Google frontier model released on October 28, 2026, positioned as the successor to Gemini 4 Argon. It accepts text, image, audio, video and file inputs such as PDFs and returns text, with a context window of 1 million tokens. Reasoning effort can be set to low, medium or high, and output can extend up to 1 million tokens through continuation. It is available through Google AI Studio and Google Vertex, and on OpenRouter via an OpenAI-compatible API.

The model is built for demanding, long-running work rather than everyday general use. It targets long-horizon agentic workflows that span multi-day projects, complex root-cause debugging, architectural design, dense multimodal synthesis, scientific proofs, security auditing and advanced coding and full-stack web development. Google also highlights state-of-the-art cybersecurity capabilities, and the model was first offered to cyber defenders through the Fairwind program.

On the Artificial Analysis Intelligence Index V5 it scores 59 points and ranks #1, one point ahead of Claude Opus 5.5, leading four of the nine evaluations in that suite. It also ranks #1 on Legal Research Bench at 55.77% and #2 out of 29 models on Vals Index at 70.36%. Its hallucination rate is 11%, higher than Gemini 4 Argon at 4%, but lower than Claude Opus 5.5 and GPT-6 Astra.

Limited timeFree trial
Open Playground ↗Create API Key ↗Getting started →

Updated:

MODEL ID

gemini-4-carbon:batch
Context
1.049M
Max output
943.718K
Recommended API
messages
Input
audio / file / image / text / video
Output
text

Highlights

  • Multimodal input: text, image, audio, video, files
  • 1 million token context window
  • Adjustable reasoning effort: low, medium, high
  • Ranked #1 on Artificial Analysis Intelligence Index V5
  • Built for multi-day agentic workflows
  • State-of-the-art cybersecurity capabilities

Providers

Best price
1 providers

Compare API prices across providers. Select a provider to view all charges and conditions.

Provider · Pricing details
Uptime
$2$1$10$5$0.1$0.05$0.1$0.05

What $10 buys

Input
5M tokens
Output
1M tokens

Estimated from current input and output prices, before discounts.

Benchmarks

  • Artificial Analysis Intelligence Index V5

    points; ranked #1

    59

  • AA-Omniscience Index

    points; highest of any model

    54

  • AA-Omniscience accuracy

    58%

  • AA-Omniscience hallucination rate

    11%

  • Vals Index

    rank #2 out of 29 models

    70.36%

  • Legal Research Bench

    rank #1

    55.77%

  • Terminal-Bench 4.0

    rank #4

    63.13%

  • SEC-bench Pro

    vulnerability discovery

    88.2%

Scores as published by the model provider.

Reasoning

Default effort
medium
Always enabled
Yes
Available effort levels
high / medium / low

Quick Start

Drop-in code to call this model with Felo API Platform compatible APIs. Start in the Playground when you want to test prompts before writing code.

Open PlaygroundCreate API Key
1

Get your API key

Create a key from your Felo API Platform account and set it as an environment variable.

export FELO_API_KEY="YOUR_API_KEY"
2

Make your first request

Use this model ID with the Felo API Platform endpoint. Optional request headers such as X-Request-Id can help troubleshooting.

gemini-4-carbon:batch

This model is reasoning-capable. If your selected protocol and model route expose reasoning controls, preserve the returned reasoning metadata when continuing a conversation.

const response = await fetch("https://openapi.felo.ai/api/v1/messages", {
  method: 'POST',
  headers: {
    Authorization: `Bearer ${process.env.FELO_API_KEY}`,
    'Content-Type': 'application/json',
  },
  body: JSON.stringify({
  "model": "gemini-4-carbon:batch",
  "max_tokens": 1024,
  "messages": [
    {
      "role": "user",
      "content": "Hello"
    }
  ]
}),
});
if (!response.ok) throw new Error(await response.text());
console.log(await response.json());

Using third-party SDKs

Use compatible SDKs by setting baseURL to https://openapi.felo.ai/api/v1. Use the Anthropic-compatible base URL for Claude-style clients. https://openapi.felo.ai/api.

3

Enable streaming

Add stream: true to receive responses as server-sent events.

curl -N https://openapi.felo.ai/api/v1/messages \
  -H "Authorization: Bearer $FELO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "gemini-4-carbon:batch",
  "max_tokens": 1024,
  "messages": [
    {
      "role": "user",
      "content": "Hello"
    }
  ],
  "stream": true
}'

Endpoint

Felo API Platform exposes compatible model-calling surfaces. Pick the surface your app or agent client already supports.

SurfaceMethodURL

Messages

Anthropic-compatible messages for Claude-style clients.

POST

https://openapi.felo.ai/api/v1/messages

Parameters

NameTypeDefaultDescription
modelstringgemini-4-carbon:batchPublic model ID to call.
max_tokensinteger—Maximum generated tokens.
messagesarray—Conversation messages for Chat Completions.
seed———
stop———
toolsarray—Tool calling definitions following compatible schema.
top_p———
reasoning———
temperaturenumber—Controls response variety when supported.
tool_choicestring or object—Controls whether a tool should be called.
response_formatobject—Request structured output when supported.
reasoning_effort———
include_reasoning———
structured_outputs———

Frequently asked questions

What is Gemini 4 Carbon (batch)?

Gemini 4 Carbon is a Google frontier model released on October 28, 2026, positioned as the successor to Gemini 4 Argon. It accepts text, image, audio, video and file inputs such as PDFs and returns text, with a context window of 1 million tokens. Reasoning effort can be set to low, medium or high, and output can extend up to 1 million tokens through continuation. It is available through Google AI Studio and Google Vertex, and on OpenRouter via an OpenAI-compatible API. The model is built for demanding, long-running work rather than everyday general use. It targets long-horizon agentic workflows that span multi-day projects, complex root-cause debugging, architectural design, dense multimodal synthesis, scientific proofs, security auditing and advanced coding and full-stack web development. Google also highlights state-of-the-art cybersecurity capabilities, and the model was first offered to cyber defenders through the Fairwind program. On the Artificial Analysis Intelligence Index V5 it scores 59 points and ranks #1, one point ahead of Claude Opus 5.5, leading four of the nine evaluations in that suite. It also ranks #1 on Legal Research Bench at 55.77% and #2 out of 29 models on Vals Index at 70.36%. Its hallucination rate is 11%, higher than Gemini 4 Argon at 4%, but lower than Claude Opus 5.5 and GPT-6 Astra.

How is usage billed?

Your bill depends on the provider you use, its prices and your usage. Longer inputs or other conditions may have different prices.

What inputs does Gemini 4 Carbon accept?

It accepts text, image, audio, video and file inputs such as PDFs, and produces text output.

How large are its context and output windows?

The context window is 1 million tokens, and output can reach up to 1 million tokens through continuation.

What is it designed for?

It targets long-horizon agentic workflows spanning multi-day projects, complex root-cause debugging, architectural design, security auditing, scientific proofs and advanced coding.

How does it differ from Gemini 4 Argon?

It is positioned as the successor to Gemini 4 Argon. Its hallucination rate is 11%, higher than Gemini 4 Argon at 4%.

Where can it be accessed?

It is available through Google AI Studio and Google Vertex, and on OpenRouter with an OpenAI-compatible API.

Related models

Gemini 4 Carbon

Googlegemini-4-carbon

Gemini 4 Carbon is Google's strongest frontier model, built on Gemini 4 Argon for long-horizon agentic workflows, cybersecurity, scientific reasoning, and complex knowledge work.

Cybersecurity vulnerability discovery and patchingLong-horizon agentic workflowsLarge-scale codebase migrationsScientific reasoning and research

Context

1.049M

Max output

943.718K

Input tokens

$4 / 1M tokens

Output tokens

$20 / 1M tokens

Gemini 4 Argon

Googlegemini-4-argon

Google's Gemini 4 Argon is a frontier model with a 1,048,576-token context window, multimodal input support, and top scores on 14 of 18 internal benchmarks.

Software engineering and codingFinance ledger reconciliationLegal case law reviewCybersecurity defense automation

Context

1.049M

Max output

943.718K

Input tokens

$4 / 1M tokens

Output tokens

$20 / 1M tokens

Gemini 3.8 Flash

Googlegemini-3.8-flash

Gemini 3.8 Flash is Google's most intelligent Flash model, a reasoning model with 1,000,000-token context and multimodal input for long-horizon agents and enterprise workflows.

Long-horizon software engineeringAutonomous agentsComplex enterprise workflowsMulti-step orchestration

Context

1.049M

Max output

65.536K

Input tokens

$1.5 / 1M tokens

Output tokens

$7.5 / 1M tokens

Gemma 4 26B A4B (free)

Googlegemma-4-26b-a4b-it:free

Gemma 4 26B A4B is an open-weight multimodal Mixture-of-Experts model from Google DeepMind with 26B total and 4B active parameters, a 256K token context, and support for text, image, and video input.

Coding assistance in IDEsAgentic workflows and tool useScientific and mathematical problem solvingDocument parsing and OCR

Context

262.144K

Max output

32.768K

Input tokens

$0 / 1M tokens

Output tokens

$0 / 1M tokens

Gemini 2.5 Pro

Googlegemini-2.5-pro

Google DeepMind's Gemini 2.5 Pro is a hybrid-reasoning, multimodal model with a 1 million token context window, built for advanced coding, agentic applications, and video understanding.

Complex multi-round reasoning tasksWeb development and web appsBuilding agentic applicationsCreative video work

Context

1.049M

Max output

65.536K

Input tokens · Base rate

$1.25 / 1M tokens

Output tokens · Base rate

$10 / 1M tokens