G
GOOGLE / gemini-4-carbon

Gemini 4 Carbon

Google multimodal reasoning model generally available since November 10. It processes text, images, video, audio and PDFs and returns text for coding, analysis and multistep tasks.

01 / SUMMARY

The essential

ORIENTATIONMultimodal reasoning and coding

Primary use stated or inferred cautiously from official documentation.

CONTEXT1,048,576 tokens shared by input and output

Maximum input capacity when published by the source.

EXIT65,536 synchronous tokens, up to 1,048,576 with long decode

Documented maximum generation limit.

ACCESSGemini API · general availability

Verified availability channels.

MODEL TYPEText and reasoning · Code · Multimodal

Primary functional family and verified specialties.

LICENSEProprietary

Declared terms for API access, model weights, or self-hosting.

02 / TECHNICAL SHEET

Limits and integration

API IDgemini-4-carbon
Model typeText and reasoning · Code · Multimodal
Access modelPaid proprietary
LicenseProprietary
DeploymentHosted API
Release10 NOV 2026
Knowledge cutoffMarch 2026
EntranceText · Image · Audio · Video · PDF
ExitText
Context window1,048,576 tokens shared by input and output
maximum output65,536 synchronous tokens, up to 1,048,576 with long decode
ReasoningNot published
Published tools
Structured OutingsNot published
Batch processingNot published
Prompt cacheNot published
Fine-tuningNot published
Verified platformsGemini API · Google AI Studio
BEST SUITED FOR

Analyze extensive documentation and code, reason about charts or combine information across formats. The API offers tools and structured outputs. The total context includes input and generation.

WORTH MONITORING

The default synchronous output limit is 65,536 tokens. Reaching one million tokens requires long decode and shares the budget with input. The card places the knowledge cutoff in March 2026.

03 / CAPABILITIES

What it can do

01

Orientation

Analyze extensive documentation and code, reason about charts or combine information across formats. The API offers tools and structured outputs. The total context includes input and generation.

02

Context

1,048,576 tokens shared by input and output · 65,536 synchronous tokens, up to 1,048,576 with long decode

03

Tools and integration

Not published

04

Access

Gemini API · general availability

Input modalities
TextImageAudioVideoPDF
04 / PRICES

Documented cost

ConceptWorthUnit/condition
Standard input$2 / 1M tokensStandard API rate
Cached input$0.10 / 1M tokensReading reused prefixes
Cache write or storageNot publishedThe condition varies by provider
Standard output$10 / 1M tokensMay include reasoning tokens
Batch input$1 / 1M tokensAsynchronous processing
Batch output$5 / 1M tokensAsynchronous processing

Through November 30, standard rates are $2 for input and $10 for output per million tokens. From December 1 they rise to $4 and $20. Batch costs half as much. Reasoning tokens are billed.

i Prices change and may depend on level, region or context length. Check the source before making a decision.

05 / EVALUATIONS

How to read the results

“

Google reports 61.2% on FrontierSWE 2 and 65.2% on Terminal-Bench 4.0 under the conditions in its model card. These results apply to the full model and the documented evaluation harnesses.

BenchmarkResultMetricSource
FrontierSWE 261,2 %Published resultView source ↗
Terminal-Bench 4.065,2 %Published resultView source ↗

Inferama only highlights a “best result” when the metric, test set, configuration, and date allow for an equivalent comparison. The supplier's figures are presented as claims from its own source.

06 / VERSIONS

Chronology

gemini-4-carbon

General availability in Gemini API.

07 / SOURCES

Traceability