G
GOOGLE / gemini-decision-1

Gemini Decision 1

Google model specialized in binary decisions, option selection and rubric scores. It evaluates text, images, audio, video and PDFs and returns probabilities and confidence without generating text explanations.

01 / SUMMARY

The essential

ORIENTATIONClassification and probabilities

Primary use stated or inferred cautiously from official documentation.

CONTEXT262,144 tokens shared by input and questions

Maximum input capacity when published by the source.

EXITUp to 1,000 typed answers; zero output tokens

Documented maximum generation limit.

ACCESSGemini Developer API · Decisions API v1beta

Verified availability channels.

MODEL TYPEMultimodal

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-decision-1
Model typeMultimodal
Access modelAPI with free and paid tiers
LicenseProprietary
DeploymentHosted API
Release11 NOV 2026
Knowledge cutoffNot published
EntranceText · Image · Audio · Video · PDF
ExitProbabilities and typed values
Context window262,144 tokens shared by input and questions
maximum outputUp to 1,000 typed answers; zero output tokens
ReasoningNot applicable
Published tools
Structured Outingsnoul / choice / score
Batch processingUnavailable
Prompt cacheUnavailable
Fine-tuningNot published
Verified platformsGemini API · Google AI Studio
BEST SUITED FOR

Classify requests, route tasks and score candidates with independent questions. One request shares input across up to 1,000 questions. Each choice or rubric accepts two to 255 options or levels.

WORTH MONITORING

No text generation, tool calls, Batch, caching, Flex or Priority. Questions accept text only. The total context includes the shared input and every question; probabilities are not guarantees of correctness.

03 / CAPABILITIES

What it can do

01

Orientation

Classify requests, route tasks and score candidates with independent questions. One request shares input across up to 1,000 questions. Each choice or rubric accepts two to 255 options or levels.

02

Context

262,144 tokens shared by input and questions · Up to 1,000 typed answers; zero output tokens

03

Tools and integration

Not published

04

Access

Gemini Developer API · Decisions API v1beta

Input modalities
TextImageAudioVideoPDF
04 / PRICES

Documented cost

ConceptWorthUnit/condition
Standard input$0.10 / 1M tokensStandard API rate
Cached inputNot publishedReading reused prefixes
Cache write or storageNot publishedThe condition varies by provider
Standard outputNot applicable: zero output tokensMay include reasoning tokens
Batch inputNot publishedAsynchronous processing
Batch outputNot publishedAsynchronous processing

Paid usage costs $0.10 per million input tokens for every modality. Shared input is billed once; each question adds its instructions and criteria. The free tier has separate quotas. There are no output-token charges.

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

“

The documentation describes calibrated decisions and confidence but publishes no model-specific benchmark table. Assess quality with your own data and thresholds suited to each task.

i We have not found published benchmarks for this model.

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-decision-1

Decisions API launch for typed multimodal decisions.

07 / SOURCES

Traceability