Back to Models

GPT Image 1 Mini

openaiopenai/gpt-image-1-mini

Cost-efficient version of GPT Image 1.

GPT Image 1 Mini (gpt-image-1-mini) is a cost-optimized variant of OpenAI's flagship image generation model, introduced at DevDay 2025 (October 6, 2025). It shares the same natively multimodal autoregressive architecture as GPT Image 1 but is tuned for speed and efficiency β€” delivering approximately 80% lower cost per image while retaining the core capabilities of text-to-image generation, image editing, and accurate text rendering. The model is ideal for high-volume production workflows where throughput and budget matter more than peak fidelity.

Through TheRouter, gpt-image-1-mini is accessible via the standard OpenAI-compatible Images API endpoint at api.therouter.ai/v1/images/generations. It supports the same feature set as the full-sized model β€” prompt-based generation, image editing, multiple quality tiers (low, medium, high), transparent-background output, and various aspect ratios. This makes it a drop-in replacement for gpt-image-1 in applications where cost sensitivity is high and the quality trade-off is acceptable.

Best for
  • β€’ High-volume image generation β€” producing thousands of images per day for content pipelines, social media assets, and marketing collateral
  • β€’ Rapid prototyping and design iteration β€” quick visual exploration during early-stage design workflows where speed outweighs final polish
  • β€’ Thumbnail and preview generation β€” creating low-to-medium quality previews, listing images, and thumbnail assets at scale
  • β€’ Real-time creative tools β€” interactive image generation in user-facing products where generation speed directly impacts UX
  • β€’ Cost-sensitive batch operations β€” bulk image creation where the 80% cost reduction makes per-request generation economically viable
Reach for something else if
  • β€’ Maximum-fidelity hero images β€” use gpt-image-1 (full) or gpt-image-2 for high-resolution marketing hero images where every detail matters
  • β€’ Complex compositional scenes with intricate prompt adherence β€” the full model handles multi-subject, detail-heavy prompts with better consistency
  • β€’ Production print assets requiring the highest resolution output β€” Mini's medium quality may not meet print-grade DPI requirements
  • β€’ Fully open-source or offline deployment β€” like all OpenAI image models, gpt-image-1-mini is proprietary and only available via API
Context Length
--
Max Output
--
Image Priceper 1M tokens
$0.0184/ image
Input Priceper 1M tokens
$2.16/ 1M tokens

Modalities

textimage→image

Capabilities

VisionImage GenerationImage Edit

Media Generation Capabilities

image_generation
sizes
  • 1024x1024
  • 1024x1536
  • 1536x1024
qualities
  • low
  • medium
  • high
max_outputs_per_request
4
output_formats
  • png
  • jpeg
  • webp
defaults
size
1024x1024
quality
medium

Pricing Breakdown

TypeRate
Image$0.0184 / image
Input$2.16 / 1M tokens
Output$10.80 / 1M tokens
Image input$2.70 / 1M tokens
Low quality$0.0054 / image
Medium quality$0.0184 / image
High quality$0.0713 / image

image is medium-quality generation price per image

Supported Parameters

promptsizequalitybackgroundoutput_formatnuser

Specifications

Release date2025-10-06openai.com β†—verified
ArchitectureAutoregressive multimodal decoder (GPT Image family), cost-optimized variantdevelopers.openai.com β†—to verify
Available output resolutions1024Γ—1024, 1024Γ—1536, 1536Γ—1024, 512Γ—512, 512Γ—768, 768Γ—512, 256Γ—256developers.openai.com β†—verified
Quality tiersLow, Medium, High (default: High for 1024+ sizes, Low for smaller sizes)developers.openai.com β†—verified
Supported output formatsPNG, JPEG, WebP (with optional transparent background)developers.openai.com β†—verified
Content moderationBuilt-in safety filter with C2PA provenance metadatadevelopers.openai.com β†—verified
LicenseProprietary (API-only, no model weights distribution)en.wikipedia.org β†—verified

Benchmarks

BenchmarkDistributionScoreSource
Cost reduction vs GPT Image 1
Low-quality images cost ~$0.005, medium-quality ~$0.017, high-quality ~$0.066 per image at 1024Γ—1024
~80% lower per image%developers.openai.com β†—
Generation speed
Measured on a typical API connection; actual speed varies by network, quality setting, and region
~20s for 1024Γ—1024 low qualitysimonwillison.net β†—
Text rendering accuracy
The mini variant retains the autoregressive decoder architecture that makes text inside images readable, though accuracy degrades slightly on very small fonts vs the full model
Comparable to GPT Image 1community.openai.com β†—

API Usage Examples

Use the global api.therouter.ai endpoint shown below for new integrations; the legacy China accelerated endpoint is retired.

Recommended: use the async API

Image generation typically takes 30–180s, beyond the edge sync timeout. The examples below use the ?async=true submit + poll pattern. Read the full async image generation & edit guide β†’

cURL
# 1) Submit job (returns 202 immediately with a polling URL).
# Image generation takes 30-180s β€” always use the async path in production.
JOB=$(curl -s -X POST "https://api.therouter.ai/v1/images/generations?async=true"   -H "Content-Type: application/json"   -H "Authorization: Bearer $THE_ROUTER_API_KEY"   -d '{
    "model": "openai/gpt-image-1-mini",
    "prompt": "A cinematic product render with soft studio lighting"
  }' | python3 -c "import sys,json;print(json.load(sys.stdin)['id'])")
echo "submitted: $JOB"

# 2) Poll until terminal (succeeded / failed / cancelled / expired).
while :; do
  R=$(curl -s "https://api.therouter.ai/v1/jobs/$JOB"     -H "Authorization: Bearer $THE_ROUTER_API_KEY")
  S=$(echo "$R" | python3 -c "import sys,json;print(json.load(sys.stdin)['status'])")
  echo "status: $S"
  case "$S" in
    succeeded) echo "$R" | python3 -c "import sys,json;print(json.load(sys.stdin)['unsigned_urls'][0])"; break ;;
    failed|cancelled|expired) echo "$R"; exit 1 ;;
  esac
  sleep 5
done

Image generation

GPT Image 1 Mini uses the Images API (not chat completions). Send a POST request to /v1/images/generations with your prompt and model ID.

cURL
curl https://api.therouter.ai/v1/images/generations \
  -H "Authorization: Bearer $THEROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-image-1-mini",
    "prompt": "A serene Japanese garden in spring with cherry blossoms, digital art style",
    "n": 1,
    "size": "1024x1024",
    "quality": "medium",
    "output_format": "png"
  }'

Image Editing Examples

Upload an image and describe the edit you want with a text prompt; the model returns the edited image as base64.

cURL
# Same async submit + poll pattern as /v1/images/generations.
JOB=$(curl -s -X POST "https://api.therouter.ai/v1/images/edits?async=true"   -H "Authorization: Bearer $THE_ROUTER_API_KEY"   -F "model=openai/gpt-image-1-mini"   -F "prompt=Turn this scene into a watercolor painting"   -F "size=1024x1024"   -F "image=@input.png" | python3 -c "import sys,json;print(json.load(sys.stdin)['id'])")
echo "submitted: $JOB"

while :; do
  R=$(curl -s "https://api.therouter.ai/v1/jobs/$JOB"     -H "Authorization: Bearer $THE_ROUTER_API_KEY")
  S=$(echo "$R" | python3 -c "import sys,json;print(json.load(sys.stdin)['status'])")
  echo "status: $S"
  case "$S" in
    succeeded) echo "$R" | python3 -c "import sys,json;print(json.load(sys.stdin)['unsigned_urls'][0])"; break ;;
    failed|cancelled|expired) echo "$R"; exit 1 ;;
  esac
  sleep 5
done

More from openai

Similar models

Cross-provider sibling models

News & changes

2025-10-06

OpenAI launches gpt-image-1-mini at DevDay 2025

OpenAI unveiled gpt-image-1-mini as a cost-optimized image generation model at DevDay 2025. The model delivers approximately 80% cost reduction compared to the full GPT Image 1 while supporting the same feature set β€” text-to-image, image editing, text rendering, and multiple output formats. The launch was part of a broader DevDay announcement that included GPT-5 Pro, Sora 2, AgentKit, and Apps SDK.

re-authored by TheRoutercommunity.openai.com β†—

Frequently asked

How does gpt-image-1-mini compare to full gpt-image-1?

GPT Image 1 Mini is approximately 80% cheaper per image while sharing the same fundamental architecture. The trade-off is in generation fidelity β€” the mini variant produces slightly less detailed outputs, particularly for complex compositional scenes, fine text rendering at small sizes, and high-resolution print-quality assets. For most web-oriented and social media use cases, the quality difference is negligible.

re-authored by TheRouterdevelopers.openai.com β†—
What quality settings are available and which should I use?

Three quality tiers are supported: 'low', 'medium', and 'high'. Low quality (~$0.005/image at 1024Γ—1024) is ideal for thumbnails, previews, and rapid prototyping. Medium quality (~$0.017/image) works well for social media posts, blog images, and most web content. High quality (~$0.066/image) is best for hero images, marketing materials, and any use case where visual impact is critical. The default quality is 'high' for 1024+ sizes and 'low' for smaller sizes.

re-authored by TheRouterdevelopers.openai.com β†—
Can I use gpt-image-1-mini via the chat completions API?

No. GPT Image 1 Mini uses the dedicated /v1/images/generations endpoint, not the chat completions endpoint. Send POST requests to api.therouter.ai/v1/images/generations with the model field set to "openai/gpt-image-1-mini". The model supports all standard OpenAI image generation parameters including prompt, size, n, quality, output_format, and background.

re-authored by TheRouterdevelopers.openai.com β†—
Fact ledger β€” every claim on this page traces here
sourceURLretrieved
Release dateopenai.com β†—2026-05-29verified
Architecturedevelopers.openai.com β†—2026-05-29to verify
Available output resolutionsdevelopers.openai.com β†—2026-05-29verified
Quality tiersdevelopers.openai.com β†—2026-05-29verified
Supported output formatsdevelopers.openai.com β†—2026-05-29verified
Content moderationdevelopers.openai.com β†—2026-05-29verified
Licenseen.wikipedia.org β†—2026-05-29verified
Cost reduction vs GPT Image 1developers.openai.com β†—2026-05-29verified
Generation speedsimonwillison.net β†—2026-05-29to verify
Text rendering accuracycommunity.openai.com β†—2026-05-29to verify
OpenAI launches gpt-image-1-mini at DevDay 2025community.openai.com β†—2026-05-29verified
How does gpt-image-1-mini compare to full gpt-image-1?developers.openai.com β†—2026-05-29to verify
What quality settings are available and which should I use?developers.openai.com β†—2026-05-29to verify
Can I use gpt-image-1-mini via the chat completions API?developers.openai.com β†—2026-05-29to verify
Help & contact