# HiDream I1 Full

> HiDream AI's 17B parameter text-to-image model using sparse diffusion transformer with mixture of experts, achieving state-of-the-art image generation quality with strong prompt following

## Quick Reference

- Model ID: hidream-i1-full
- Creator: HiDream AI
- Status: active
- Family: hidream
- Base URL: https://api.lumenfall.ai/openai/v1

## Specifications

- Input Modalities: text, image
- Output Modalities: image
- Supported Modes: Text to Image, Image Edit

## API Parameters

The compiled parameter schema for this model is available via the API: `GET /v1/models/hidream-i1-full?schema=true`.

### Core Parameters

- `prompt` (string) — REQUIRED: Edit instruction for the image. Modes: Image Edit, Text to Image
- `negative_prompt` (string): Negative prompt to guide generation away from undesired content. Modes: Image Edit, Text to Image
- `seed` (integer): Random seed for reproducibility. Modes: Image Edit, Text to Image

### Size & Layout

- `size` (string): Image dimensions as WxH pixels (e.g. "1024x1024") or aspect ratio (e.g. "16:9"). Modes: Image Edit, Text to Image
- `aspect_ratio` (string): Aspect ratio of the output image (e.g. "16:9", "1:1"). Modes: Image Edit, Text to Image
- `resolution` (string): Output resolution tier (e.g. "1K", "4K"). Modes: Image Edit, Text to Image

### Media Inputs

- `image` (file) — REQUIRED: Input image(s) to edit. Modes: Image Edit

### Output & Format

- `response_format` (string): How to return the image. Default: url. Values: url, b64_json. Modes: Image Edit, Text to Image
- `output_format` (string): Output image format. Values: png, jpeg, gif, webp, avif. Modes: Image Edit, Text to Image
- `output_compression` (integer): Compression level for lossy formats (JPEG, WebP, AVIF). Modes: Image Edit, Text to Image
- `n` (integer): Number of images to generate. Default: 1. Modes: Image Edit, Text to Image

### Additional Parameters

- `cfg_scale` (number): Classifier-free guidance scale — higher values stick more closely to the prompt. Modes: Image Edit, Text to Image
- `strength` (number): How much to transform the input image: 0 keeps it unchanged, 1 fully regenerates from the prompt. Modes: Image Edit
- `enable_safety_checker` (boolean): If set to true, the safety checker will be enabled.. Modes: Image Edit, Text to Image. Only available via fal
- `loras` (array): A list of LoRAs to apply to the model. Each LoRA specifies its path, scale, and optional weight name.. Modes: Image Edit, Text to Image. Only available via fal
- `num_inference_steps` (integer): The number of inference steps to perform.. Modes: Image Edit, Text to Image. Only available via fal
- `sync_mode` (boolean): If `True`, the media will be returned as a data URI and the output data won't be available in the request history.. Modes: Image Edit, Text to Image. Only available via fal


## Model Identifiers

- Primary Slug: hidream-i1-full

## Dates

- Released: April 2025

## Tags

image-generation, text-to-image

## Available Providers

### fal.ai

- Config Key: fal/hidream-i1-full-edit
- Provider Model ID: fal-ai/hidream-i1-full/image-to-image
- Pricing: $0.050/megapixel
  - Source: https://fal.ai/models/fal-ai/hidream-i1-full

### fal.ai

- Config Key: fal/hidream-i1-full
- Provider Model ID: fal-ai/hidream-i1-full
- Pricing: $0.050/megapixel
  - Source: https://fal.ai/models/fal-ai/hidream-i1-full


## Arena Benchmarks

### The Reversed Rodeo

- Elo: 1137
- Record: 2W / 3L / 0T (5 battles)
- Rank: #9 of 68

### The Halloween Invitation

- Elo: 1133
- Record: 2W / 2L / 0T (4 battles)
- Rank: #12 of 70


## Use Cases & Category Performance

### Art (Text-to-Image)

- Rank: #13 of 31
- Elo: 1176
- Record: 4W / 5L / 0T (9 battles)
- Win Rate: 44.4%

### Creativity (Text-to-Image)

- Rank: #46 of 62
- Elo: 1158
- Record: 4W / 7L / 0T (11 battles)
- Win Rate: 36.4%

### Product, Branding & Commercial (Text-to-Image)

- Rank: #50 of 61
- Elo: 1143
- Record: 3W / 4L / 0T (7 battles)
- Win Rate: 42.9%

### Text Rendering (Text-to-Image)

- Rank: #52 of 61
- Elo: 1137
- Record: 3W / 4L / 0T (7 battles)
- Win Rate: 42.9%

### Photorealism (Text-to-Image)

- Rank: #55 of 62
- Elo: 1124
- Record: 4W / 10L / 1T (15 battles)
- Win Rate: 26.7%

### Aesthetics (Text-to-Image)

- Rank: #57 of 62
- Elo: 1167
- Record: 7W / 12L / 1T (20 battles)
- Win Rate: 35.0%

### Prompt Adherence (Text-to-Image)

- Rank: #58 of 61
- Elo: 1150
- Record: 4W / 7L / 1T (12 battles)
- Win Rate: 33.3%


## Image Gallery

5 images available for this model. Browse all at https://lumenfall.ai/models/hidream-ai/hidream-i1-full/gallery

### Curated Examples

- [Cinematic wide shot of a high-end, contemporary boutique storefront at golden hour. The storefron...](https://assets.lumenfall.ai/qN5s8LtmnpVZ1xdXRX3mKCWSUTTDFNlPsX9na954EdY/rs:fit:1500:1500/plain/gs://lumenfall-prod-assets/4p3opw034f68ob1m2ia1b56gpz4g@jpeg)
- [A hyper-realistic close-up of an elderly craftsman's weathered hands meticulously assembling a vi...](https://assets.lumenfall.ai/Znz5JyKKN3hg6gcyHXq5wqYP3NjpjVf1oWoGNfJTu1g/rs:fit:1500:1500/plain/gs://lumenfall-prod-assets/lum3otxgqzq6r6i9yiutfyd7rt8h@jpeg)
- [A vibrant, hand-painted wooden sign for "SOLSTICE GARDENS" hanging over a rustic flower shop. The...](https://assets.lumenfall.ai/EfzEU8_pHG98rTXKfIDMf6KeDX5TttPOBdJWn-5Y88U/rs:fit:1500:1500/plain/gs://lumenfall-prod-assets/r08hlr24z1swh9xcjx6ezciecyra@jpeg)

### Arena Competition Results

- [The Reversed Rodeo](https://assets.lumenfall.ai/Qw4fBSeUPdw-CcDWv5iwTPurd6Hf3egxrAT1KNsAVNw/rs:fit:1500:1500/plain/gs://lumenfall-prod-assets/e6n3o0clul4fj93246gclz30c4k2@jpeg): #9 of 68 (Elo 1137)
- [The Halloween Invitation](https://assets.lumenfall.ai/gTgrFPE4WnQUgvGgxmIAo3Wx86stphb6zKyiSc3D2Mw/rs:fit:1500:1500/plain/gs://lumenfall-prod-assets/8yi9lt1bgkpkdotn1l41jr86enrb@jpeg): #12 of 70 (Elo 1133)

## Example Prompt

The following prompt was used to generate an example image in our playground:

A vibrant, hand-painted wooden sign for "SOLSTICE GARDENS" hanging over a rustic flower shop. The gold lettering is elegant and perfectly rendered. In the soft-focus background, a capybara is calmly sitting among the potted clay ferns.

## Code Examples

### Text to Image (/v1/images/generations)

#### cURL

curl -X POST \
  https://api.lumenfall.ai/openai/v1/images/generations \
  -H "Authorization: Bearer $LUMENFALL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "hidream-i1-full",
    "prompt": "",
    "size": "1024x1024"
  }'

# Response:
# { "created": 1234567890, "data": [{ "url": "https://...", "revised_prompt": "..." }] }

#### JavaScript

import OpenAI from 'openai';

const client = new OpenAI({
  apiKey: 'YOUR_API_KEY',
  baseURL: 'https://api.lumenfall.ai/openai/v1'
});

const response = await client.images.generate({
  model: 'hidream-i1-full',
  prompt: '',
  size: '1024x1024'
});

// { created: 1234567890, data: [{ url: "https://...", revised_prompt: "..." }] }
console.log(response.data[0].url);

#### Python

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_API_KEY",
    base_url="https://api.lumenfall.ai/openai/v1"
)

response = client.images.generate(
    model="hidream-i1-full",
    prompt="",
    size="1024x1024"
)

# { created: 1234567890, data: [{ url: "https://...", revised_prompt: "..." }] }
print(response.data[0].url)

### Image Edit (/v1/images/edits)

#### cURL

curl -X POST \
  https://api.lumenfall.ai/openai/v1/images/edits \
  -H "Authorization: Bearer $LUMENFALL_API_KEY" \
  -F "model=hidream-i1-full" \
  -F "image=@source.png" \
  -F "prompt=Add a starry night sky to this image" \
  -F "size=1024x1024"

# Response:
# { "created": 1234567890, "data": [{ "url": "https://...", "revised_prompt": "..." }] }

#### JavaScript

import OpenAI from 'openai';
import fs from 'fs';

const client = new OpenAI({
  apiKey: 'YOUR_API_KEY',
  baseURL: 'https://api.lumenfall.ai/openai/v1'
});

const response = await client.images.edit({
  model: 'hidream-i1-full',
  image: fs.createReadStream('source.png'),
  prompt: 'Add a starry night sky to this image',
  size: '1024x1024'
});

// { created: 1234567890, data: [{ url: "https://...", revised_prompt: "..." }] }
console.log(response.data[0].url);

#### Python

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_API_KEY",
    base_url="https://api.lumenfall.ai/openai/v1"
)

response = client.images.edit(
    model="hidream-i1-full",
    image=open("source.png", "rb"),
    prompt="Add a starry night sky to this image",
    size="1024x1024"
)

# { created: 1234567890, data: [{ url: "https://...", revised_prompt: "..." }] }
print(response.data[0].url)


## About

## Overview
HiDream I1 Full is a high-capacity text-to-image and image-to-image model developed by HiDream AI. Utilizing a 17-billion parameter architecture, it is designed to bridge the gap between complex natural language prompts and high-fidelity visual outputs. The model is distinctive for its use of a Sparse Diffusion Transformer (DiT) combined with a Mixture-of-Experts (MoE) framework, allowing it to handle massive parameter counts with localized computational efficiency.

## Strengths
*   **Prompt Adherence:** The model demonstrates high fidelity to long, descriptive, and nuanced text prompts, accurately placing specific objects and attributes within a scene as requested.
*   **Compositional Detail:** Due to the large parameter count, it excels at rendering complex textures and lighting conditions that smaller diffusion models often struggle to resolve.
*   **MoE Efficiency:** The Mixture-of-Experts architecture allows the model to activate only a subset of its 17B parameters per request, leading to sophisticated image generation qualities without the prohibitive latency typical of monolithic models of this scale.
*   **Multimodal Input:** It native supports image-to-image workflows, allowing users to provide visual references to guide the style, structure, or content of the final output.

## Limitations
*   **Computational Cost:** At $0.05 per generation, the model is targeted towards high-end production use cases and may be less economical for high-volume, low-stakes applications compared to smaller distilled models.
*   **Specialized Hardware Requirements:** Due to the 17B parameter size, local deployment is challenging, making it primarily a cloud-driven model for most enterprise environments.
*   **Training Cutoff:** Like all diffusion models, it may lack specific knowledge of very recent events, niche brand logos, or specialized technical schematics unless explicitly provided in the prompt or through fine-tuning.

## Technical Background
HiDream I1 Full is built on a Diffusion Transformer (DiT) backbone, a departure from the traditional U-Net architectures used in earlier generative models. It integrates a Sparse Mixture-of-Experts (MoE) layer, which scales the model's capacity to 17 billion parameters while maintaining manageable inference speeds. This training approach emphasizes high-dimensional latent space representations to ensure that fine-grained textual details are mapped accurately to pixels.

## Best For
*   **Professional Concept Art:** Generating high-resolution environment and character designs where specific lighting and material properties are critical.
*   **Precision Marketing Assets:** Creating brand-aligned imagery that requires strict adherence to complex creative briefs.
*   **Visual Prototyping:** Rapidly iterating on product designs using image-to-image guidance to maintain structural consistency.

HiDream I1 Full is available for immediate testing and deployment through Lumenfall’s unified API and interactive playground.

## Frequently Asked Questions

### How much does HiDream I1 Full cost?

HiDream I1 Full starts at $0.05 per image through Lumenfall. Pricing varies by provider. Lumenfall does not add any markup to provider pricing.

### How do I use HiDream I1 Full via API?

You can use HiDream I1 Full through Lumenfall's OpenAI-compatible API. Send requests to the unified endpoint with model ID "hidream-i1-full". Code examples are available in Python, JavaScript, and cURL.

### Which providers offer HiDream I1 Full?

HiDream I1 Full is available through fal.ai on Lumenfall. Lumenfall automatically routes requests to the best available provider.

## Links

- Model Page: https://lumenfall.ai/models/hidream-ai/hidream-i1-full
- About: https://lumenfall.ai/models/hidream-ai/hidream-i1-full/about
- Providers, Pricing & Performance: https://lumenfall.ai/models/hidream-ai/hidream-i1-full/providers
- API Reference: https://lumenfall.ai/models/hidream-ai/hidream-i1-full/api
- Benchmarks: https://lumenfall.ai/models/hidream-ai/hidream-i1-full/benchmarks
- Use Cases: https://lumenfall.ai/models/hidream-ai/hidream-i1-full/use-cases
- Gallery: https://lumenfall.ai/models/hidream-ai/hidream-i1-full/gallery
- Playground: https://lumenfall.ai/models/hidream-ai/hidream-i1-full/playground
- API Documentation: https://docs.lumenfall.ai