Zum Inhalt springen
AI.info

Model Tracker

Gemini 4 Carbon

Gemini 4 Carbon is an AI model made by Google DeepMind. Google DeepMind released it on 10 November 2026. It takes audio, image, text, video and returns text.

Auf Englisch

Google DeepMind

Gemini 4 Carbon is an AI model made by Google DeepMind. Google DeepMind released it on 10 November 2026. It takes audio, image, text, video and returns text. Its context window is 1,048,576 tokens. Through the API it costs $2.00 per 1M tokens for input and $10.00 per 1M tokens for output.

Released
10 November 2026 — models.dev, read 20 November 2026, from public catalogs that agree (models.dev, Epoch AI), not yet checked on the maker’s page
Status
Not confirmed: The maker describes a limited rollout and pre-release testing, but does not explicitly label the model preview, beta, experimental, generally available, deprecated, or retired.
Takes
Audio, Image, Text, Video — models.dev, read 20 November 2026, from public catalogs that agree (models.dev, OpenRouter), not yet checked on the maker’s page
Returns
Text — models.dev, read 20 November 2026, from public catalogs that agree (models.dev, OpenRouter), not yet checked on the maker’s page
Open weights
No — models.dev, read 20 November 2026, from public catalogs that agree (models.dev, Epoch AI), not yet checked on the maker’s page
Licence
Not confirmed: The maker’s pages checked do not name a licence for Gemini 4 Carbon.
Context window
1,048,576 tokens — models.dev, read 20 November 2026, from public catalogs that agree (models.dev, OpenRouter), not yet checked on the maker’s page
Longest answer
Not confirmed: The maker’s pages checked do not print a standard maximum-output-token limit for Gemini 4 Carbon.
Knowledge cutoff
Not confirmed: The maker’s pages checked do not state a knowledge-cutoff date for Gemini 4 Carbon.
API price, input
$2.00 per 1M tokens — models.dev, read 20 November 2026, from public catalogs that agree (models.dev, OpenRouter, LiteLLM), not yet checked on the maker’s page
API price, output
$10.00 per 1M tokens — models.dev, read 20 November 2026, from public catalogs that agree (models.dev, OpenRouter, LiteLLM), not yet checked on the maker’s page
Where
Not confirmed: The maker says access is beginning for a subset of cyber defenders through Fairwind, but the pages checked do not confirm availability through any of the register’s listed channels: API, app, downloadable weights, or cloud marketplaces.
Parameters
Not confirmed: The maker’s pages checked do not publish a parameter count for Gemini 4 Carbon.
Model card
Not confirmed: The maker’s model-card index lists Gemini 4 Argon but does not list a Gemini 4 Carbon model card.

Some figures come from public model catalogs that agree; each is being checked on the maker’s page.

In the news

Benchmarks

Measured by others, not by the maker: each score with the name the measurer gives the model and the setting it ran, its day and its licence.

Arena

Arena’s leaderboard dataset (lmarena-ai/leaderboard-dataset on Hugging Face), CC BY 4.0: ratings from people’s votes between two anonymous models, as Arena publishes them. Data · CC BY 4.0

Epoch AI

Epoch AI, ‘Capabilities & benchmarking’, epoch.ai/benchmarks, CC BY 4.0: Epoch AI’s own runs and its Capabilities Index. Data · CC BY 4.0

Questions

When was Gemini 4 Carbon released?

Google DeepMind released Gemini 4 Carbon on 10 November 2026.

What is the context window of Gemini 4 Carbon?

The context window of Gemini 4 Carbon is 1,048,576 tokens, as Google DeepMind prints it.

How much does the Gemini 4 Carbon API cost?

On the standard tier, Gemini 4 Carbon costs $2.00 per 1M tokens for input and $10.00 per 1M tokens for output, as Google DeepMind lists it.

Are the weights of Gemini 4 Carbon open?

No: Google DeepMind does not publish the weights of Gemini 4 Carbon.

What does Gemini 4 Carbon take and return?

Gemini 4 Carbon takes audio, image, text, video and returns text.

More from Google DeepMind

Sources

  1. Gemini 4 Carbon: our next frontier model (read 30 October 2026)
  2. Model cards (read 30 October 2026)
  3. Models (read 30 October 2026)
  4. models.dev model catalog (read 20 November 2026)
  5. Epoch AI model catalog (read 20 November 2026)
  6. OpenRouter model catalog (read 20 November 2026)
  7. LiteLLM model catalog (read 20 November 2026)

All of the Model Tracker