Amazon / Titan Image Generator G1

Titan Image Generator G1 v1

Amazon image generation and editing model with text-to-image generation, inpainting, outpainting and image variations. This record preserves its original version-specific capabilities and historical pricing.

Image generation

historicalClosed weightsImage generationImage editing

Official documentation (opens in a new tab)Reviewed 14 Oct 2026
17 of 28 specification fields collected

Output resolution25 optionsSee specification notes below

Text input limit512512

Input billingSee ratesAmazon Bedrock (historical model)

Output billingSee ratesPublished units and conditions below

Specifications

Provider-documented facts. “Not collected” means we have not verified that field yet, not that the feature is unsupported.

Context / input limit: Not collected
Maximum input tokens: Not collected
Maximum output: Not collected
Reasoning: Not collected
Tool calling: Not collected
Structured output: Not collected
Streaming: Not collected
Batch processing: Not collected
Fine-tuning: SupportedSource for Fine-tuning (opens in a new tab)
Context caching: Not collected
Parameters: Not collected
Knowledge cutoff: Not collected
Input modalities: Text, ImageSource for Input modalities (opens in a new tab)
Output modalities: ImageSource for Output modalities (opens in a new tab)
Languages: EnglishSource for Languages (opens in a new tab)
Maximum prompt characters: 512Source for Maximum prompt characters (opens in a new tab)
Input file formats: JPEG, PNGSource for Input file formats (opens in a new tab)
Output formats: Base64-encoded image strings in the JSON responseSource for Output formats (opens in a new tab)
Images per request: 5Source for Images per request (opens in a new tab)
Image input limits: Variation accepts 1–5 reference images; inpainting/outpainting inputs and masks have a 1,408-pixel longer-side limit.Source for Image input limits (opens in a new tab)
Output resolutions: View all 25 entries
1024 × 1024
768 × 768
512 × 512
768 × 1152
384 × 576
1152 × 768
576 × 384
768 × 1280
384 × 640
1280 × 768
640 × 384
896 × 1152
448 × 576
1152 × 896
576 × 448
768 × 1408
384 × 704
1408 × 768
704 × 384
640 × 1408
320 × 704
1408 × 640
704 × 320
1152 × 640
1173 × 640
Source for Output resolutions (opens in a new tab)
Generation controls: Standard or premium quality; cfgScale 1.1–10 (default 8); seed 0–2,147,483,646 (default 42); variation similarityStrength 0.2–1.0 (default 0.7).Source for Generation controls (opens in a new tab)
Supported API routes: InvokeModelSource for Supported API routes (opens in a new tab)
Image and motion controls: Text-to-image; inpainting/outpainting using a text-described or binary image mask; image variations; reference-based style customization.Source for Image and motion controls (opens in a new tab)
Quality and watermark controls: Invisible watermark in generated images, with a separate watermark detection API.Source for Quality and watermark controls (opens in a new tab)
Release date: 2023-11-29Source for Release date (opens in a new tab)
Endpoint availability: Historical original version. Its current standalone deployment availability and exact shutdown date were not verified; absence from today’s model-card index is not sufficient to invent a retirement date.Source for Endpoint availability (opens in a new tab)
Customization options: Historical fine-tuning workflow accepts 5–10,000 image-caption pairs in JSONL, with images stored in S3; custom-model inference uses provisioned throughput.Source for Customization options (opens in a new tab)

Access & pricing

Prices belong to an endpoint and billing tier. A provider listing does not guarantee access for every account or region.

Amazon Bedrock (historical model)

Provider source (opens in a new tab)

API model ID: amazon.titan-image-generator-v1
Access: Historical original version. Its current standalone deployment availability and exact shutdown date were not verified; absence from today’s model-card index is not sufficient to invent a retirement date.
Regions: Original launch: us-east-1 (N. Virginia) and us-west-2 (Oregon). Historical coverage, not a guarantee of current access.

Prices retain their published currencies and billing units below.

Published billing units · original currency · observed 14 Oct 2026
Category	Usage	Rate	Unit
Historical pricing	Standard image · Smaller than 512 x 512¹	$0.008	image
Historical pricing	Premium image · Smaller than 512 x 512¹	$0.01	image
Historical pricing	Standard image · Larger than 512 x 512¹	$0.01	image
Historical pricing	Premium image · Larger than 512 x 512¹	$0.012	image
Historical pricing	Base model throughput · 1-month commitment²	$16.20	model-unit hour
Historical pricing	Base model throughput · 6-month commitment²	$13.00	model-unit hour
Historical pricing	Custom model throughput · no commitment²	$23.40	model-unit hour
Historical pricing	Custom model throughput · 1-month commitment²	$21.00	model-unit hour
Historical pricing	Custom model throughput · 6-month commitment²	$16.85	model-unit hour

Rate conditions

US East (N. Virginia) selection; published AWS price map 2026-10-14T03:30:16Z. Retained meter for a historical model, not confirmation of service availability. Literal pricing-template bucket label. Boundaries conflict with the image guide and resolution-equivalent table; see source conditions before using a particular size. No availability implied.
US East (N. Virginia) selection; published AWS price map 2026-10-14T03:30:16Z. Retained meter for a historical model, not confirmation of service availability. Separate provisioned-throughput model-unit charge, not a per-image price. Historical eligibility and any private continuation contract must be checked separately.

Pricing source (opens in a new tab)

Data & enterprise terms

Refer to the Bedrock data policy. No inference or customization request was made in this review.

Read the service policy (opens in a new tab)

Caching

Caching can reuse earlier input processing. Availability, lifetime and billing depend on the endpoint. A cache hit still generates a new response.

01 / INITIAL INPUTProcess & retain

The system processes eligible input and retains reusable state.

02 / LATER REQUESTCheck available state

Reuse depends on matching input, session state and the cache policy.

03 / CACHE HITReuse & generate

Reusing state can reduce repeated work. Any charges follow the billing rules below.

Amazon Bedrock (historical model)

amazon.titan-image-generator-v1

Not verified

No Titan-version-specific prompt-cache contract verified. Nova caching rules are not inherited.

How it works: Not collected
Minimum reusable input: Not collected
Lifetime & refresh: Not collected
How to enable it: Not collected
Usage fields: Not collected
Billing rules: Not collected
Limits & data boundaries: Not collected

No verified cache rate for this endpoint. An empty rate is not a zero charge.

Caching reviewed 14 Oct 2026

What the measurements say

No benchmark observations have been added for this model. We do not turn a vendor description into a performance score.

Latency and throughput depend on the host, request size, reasoning settings and load. There is no measured speed or global usage figure in this record.

The record behind the model

Why it is included

Amazon identifies Titan as its own foundation model family delivered through Amazon Bedrock. The reviewed original product and API documentation offer managed inference/customization, not downloadable original model weights.

Classification evidence (opens in a new tab)

Lifecycle dates

General availability and watermark detection announced23 Apr 2024Source (opens in a new tab)

Sources reviewed

amazon titan image1 ga (opens in a new tab)Reviewed 14 Oct 2026

amazon titan image1 customization (opens in a new tab)Reviewed 14 Oct 2026

amazon titan image parameters (opens in a new tab)Reviewed 14 Oct 2026

amazon bedrock document history (opens in a new tab)Reviewed 14 Oct 2026

amazon model cards current (opens in a new tab)Reviewed 14 Oct 2026

amazon bedrock pricing current (opens in a new tab)Reviewed 14 Oct 2026

amazon titan price map (opens in a new tab)Reviewed 14 Oct 2026

Model notes

24 records cover Nova, Titan text/image/embedding and reranking families. Historical generation versions include exact selectors, retained prices and explicit lifecycle uncertainty. Regional billing, older identities and several specifications remain incomplete.

Official Amazon provider icon used as a fallback; distinct model-specific artwork not verified.

This is a historical model record. Retained documentation or pricing does not establish that new deployments are available.

The overview describes pixel-area billing while the parameter table maps several larger dimensions to the 512-equivalent tier. The pricing template also labels the v2 high tier as larger than 1024 × 1024. Because those statements conflict, published bucket labels and rate values are preserved without computing a per-resolution quote.

Check against your workload

Try representative inputs, measure task success and total cost, then check the exact endpoint’s data policy.

Read our methodology