Midjourney

Access Midjourney's AI models including V6, V6.1, and V7 for image generation through Runware's unified API. Learn about Midjourney-specific parameters, quality controls, and stylization options.

Introduction

Midjourney's AI models are integrated into the Runware platform through our unified API, providing access to flagship text-to-image generation technology renowned for cinematic depth, vivid color motion, and painterly realism inspired by physical art and photography.

Through the providerSettings.midjourney object, you can access Midjourney's unique parameters for controlling quality, stylization intensity, creative variation, and artistic effects, while maintaining the consistency of Runware's standard API structure. This page documents the technical specifications, parameter requirements, and provider-specific settings for all Midjourney models available through our platform.

providerSettings » midjourney midjourney object

Configuration object for Midjourney-specific generation settings. These parameters provide fine control over quality, artistic stylization, creative variation, and experimental effects.

View example

{
  "taskType": "imageInference",
  "taskUUID": "a770f077-f413-47de-9dac-be0b26a35da6",
  "positivePrompt": "Fantasy castle on a cliff at sunset",
  "model": "midjourney:3@1",
  "width": 1456,
  "height": 816,
  "providerSettings": {
    "midjourney": {
      "quality": 2,
      "stylize": 500,
      "chaos": 25
    }
  }
}

Properties ⁨5⁩ properties

providerSettings » midjourney » quality quality "0.25" | "0.5" | "1" | "2" Default: 1

Controls how much computational effort Midjourney spends rendering the image. Higher values use more GPU time and improve fine details, texture accuracy and visual refinement, but increase generation time. This parameter does not affect final resolution or image format quality.

Available values:

0.25: Very fast generation, simplified details, minimal refinement.
0.5: Faster rendering with moderate detail quality.
1: Default setting with a balanced level of detail and speed.
2: Highest refinement and texture detail with significantly longer generation time.

Midjourney V7 only supports quality values of 1 and 2.

providerSettings » midjourney » stylize stylize integer Min: 0 Max: 1000 Default: 100

Governs how much Midjourney's artistic dataset style affects the image. Lower values follow the prompt more literally, while higher values apply more artistic interpretation and Midjourney's aesthetic training.

A value of 0 minimizes stylistic influence, while values approaching 1000 maximize artistic interpretation. Most use cases benefit from values between 50-500.

providerSettings » midjourney » chaos chaos integer Min: 0 Max: 100 Default: 0

Controls variation and unpredictability in the generation process. A value of 0 disables chaos, producing consistent results for the same prompt. Higher values create more diverse and less predictable outputs, useful for creative exploration.

Values above 50 produce highly varied results, while values between 10-30 add subtle variation without dramatically changing the core composition.

providerSettings » midjourney » weird weird integer Min: 0 Max: 3000 Default: 0: Adds surreal and experimental characteristics to the generated image. A value of 0 disables the effect, producing conventional results. Higher values introduce increasingly abstract and unusual visual elements.

Very high values (e.g., 2000-3000) produce extreme and often impractical results. Most creative use cases work best with values between 100-1000.

providerSettings » midjourney » niji niji "0" | "5" | "6" | "close" Default: close

Selects the rendering engine, with specific support for anime-style generation through the Niji models.

Available values:

close: Standard Midjourney rendering (default).
0: Disable Niji mode.
5: Niji 5 anime-style rendering.
6: Niji 6 anime-style rendering (latest).

The Niji engines are optimized for anime and manga aesthetics, producing results with characteristic anime visual styles and techniques.

Image models

Midjourney V6

Flagship text-to-image model by Midjourney Inc. delivering cinematic depth, vivid color motion, and painterly realism inspired by physical art and photography. This model offers a major leap in detail, structure, and prompt accuracy with strong control over composition and style.

Model AIR ID: midjourney:1@1.

Supported workflows: Text-to-image, image-to-image.

Technical specifications:

Positive prompt: 1-2000 characters.
Number of results: Must be a multiple of 4 (4, 8, 12, 16, 20, Default: 4).
Reference images: Supports 1 image via inputs.referenceImages.
Supported dimensions: 1456×816 (16:9), 816×1456 (9:16), 1024×1024 (1:1), 1232×928 (4:3), 928×1232 (3:4), 1344×896 (3:2), 896×1344 (2:3), 1680×720 (21:9).

Provider-specific settings:

Parameters supported: quality, stylize, chaos, weird, niji.

Standard generation

{
  "taskType": "imageInference",
  "taskUUID": "f47ac10b-58cc-4372-a567-0e02b2c3d492",
  "model": "midjourney:1@1",
  "positivePrompt": "Fantasy landscape with ancient ruins, dramatic lighting, concept art style",
  "width": 1456,
  "height": 816
}

High stylization

{
  "taskType": "imageInference",
  "taskUUID": "6ba7b829-9dad-11d1-80b4-00c04fd430c8",
  "model": "midjourney:1@1",
  "positivePrompt": "Cyberpunk cityscape at night, neon lights reflecting on wet streets",
  "width": 1024,
  "height": 1024,
  "providerSettings": {
    "midjourney": {
      "quality": 2,
      "stylize": 750
    }
  }
}

Experimental variation

{
  "taskType": "imageInference",
  "taskUUID": "550e8400-e29b-41d4-a716-446655440010",
  "model": "midjourney:1@1",
  "positivePrompt": "Abstract portrait with geometric patterns and bold colors",
  "width": 928,
  "height": 1232,
  "providerSettings": {
    "midjourney": {
      "chaos": 50,
      "weird": 500
    }
  }
}

Image-to-image with reference

{
  "taskType": "imageInference",
  "taskUUID": "6ba7b829-9dad-11d1-80b4-00c04fd430c8",
  "model": "midjourney:1@1",
  "positivePrompt": "Fantasy character in the same art style",
  "inputs": {
    "referenceImages": ["c64351d5-4c59-42f7-95e1-eace013eddab"]
  },
  "width": 928,
  "height": 1232
}

Midjourney V6.1

Refined version of V6 offering improved lighting behavior, spatial coherence, and tonal balance for more natural, cinematic compositions. This model includes sharper details, better anatomy and textures, reduced artifacts, and improved text readability.

Model AIR ID: midjourney:2@1.

Supported workflows: Text-to-image, image-to-image.

Technical specifications:

Positive prompt: 1-2000 characters.
Number of results: Must be a multiple of 4 (4, 8, 12, 16, 20, Default: 4).
Reference images: Supports 1 image via inputs.referenceImages.
Supported dimensions: 1456×816 (16:9), 816×1456 (9:16), 1024×1024 (1:1), 1232×928 (4:3), 928×1232 (3:4), 1344×896 (3:2), 896×1344 (2:3), 1680×720 (21:9).

Provider-specific settings:

Parameters supported: quality, stylize, chaos, weird, niji.

Product visualization

{
  "taskType": "imageInference",
  "taskUUID": "f47ac10b-58cc-4372-a567-0e02b2c3d493",
  "model": "midjourney:2@1",
  "positivePrompt": "Luxury watch on marble surface, studio lighting, product photography",
  "width": 1232,
  "height": 928,
  "providerSettings": {
    "midjourney": {
      "quality": 2,
      "stylize": 300
    }
  }
}

Character art

{
  "taskType": "imageInference",
  "taskUUID": "6ba7b830-9dad-11d1-80b4-00c04fd430c8",
  "model": "midjourney:2@1",
  "positivePrompt": "Fantasy warrior character, detailed armor and weapons, dramatic pose",
  "width": 816,
  "height": 1456,
  "providerSettings": {
    "midjourney": {
      "quality": 1,
      "stylize": 500
    }
  }
}

Character art with reference

{
  "taskType": "imageInference",
  "taskUUID": "6ba7b830-9dad-11d1-80b4-00c04fd430c8",
  "model": "midjourney:2@1",
  "positivePrompt": "Fantasy warrior character with detailed armor",
  "inputs": {
    "referenceImages": ["c64351d5-4c59-42f7-95e1-eace013eddab"]
  },
  "width": 816,
  "height": 1456,
  "providerSettings": {
    "midjourney": {
      "quality": 1,
      "stylize": 500
    }
  }
}

Midjourney V7

Next-generation Midjourney release emphasizing realism and control. Produces grand, cinematic imagery with fluid brushwork and deeply atmospheric tone. Features significant improvements in visual realism, texture fidelity, lighting, natural-language understanding, and photographic quality.

Model AIR ID: midjourney:3@1.

Supported workflows: Text-to-image, image-to-image.

Technical specifications:

Positive prompt: 1-2000 characters.
Number of results: Must be a multiple of 4 (4, 8, 12, 16, 20, Default: 4).
Reference images: Supports 1 image via inputs.referenceImages.
Supported dimensions: 1456×816 (16:9), 816×1456 (9:16), 1024×1024 (1:1), 1232×928 (4:3), 928×1232 (3:4), 1344×896 (3:2), 896×1344 (2:3), 1680×720 (21:9).

Provider-specific settings:

Parameters supported: quality, stylize, chaos, weird, niji.

Photorealistic portrait

{
  "taskType": "imageInference",
  "taskUUID": "f47ac10b-58cc-4372-a567-0e02b2c3d494",
  "model": "midjourney:3@1",
  "positivePrompt": "Portrait of a woman in natural light, editorial photography style, shallow depth of field",
  "width": 816,
  "height": 1456,
  "providerSettings": {
    "midjourney": {
      "quality": 2,
      "stylize": 200
    }
  }
}

Architectural visualization

{
  "taskType": "imageInference",
  "taskUUID": "6ba7b831-9dad-11d1-80b4-00c04fd430c8",
  "model": "midjourney:3@1",
  "positivePrompt": "Modern minimalist house with floor-to-ceiling windows, golden hour lighting, architectural photography",
  "width": 1456,
  "height": 816,
  "providerSettings": {
    "midjourney": {
      "quality": 2,
      "stylize": 100
    }
  }
}

Cinematic landscape

{
  "taskType": "imageInference",
  "taskUUID": "550e8400-e29b-41d4-a716-446655440011",
  "model": "midjourney:3@1",
  "positivePrompt": "Misty mountain range at dawn, dramatic volumetric lighting, cinematic composition",
  "width": 1680,
  "height": 720,
  "providerSettings": {
    "midjourney": {
      "quality": 2,
      "stylize": 150
    }
  }
}

Cinematic landscape with reference

{
  "taskType": "imageInference",
  "taskUUID": "550e8400-e29b-41d4-a716-446655440011",
  "model": "midjourney:3@1",
  "positivePrompt": "Misty mountain range at dawn with similar atmosphere and color grading",
  "inputs": {
    "referenceImages": ["c64351d5-4c59-42f7-95e1-eace013eddab"]
  },
  "width": 1680,
  "height": 720,
  "providerSettings": {
    "midjourney": {
      "quality": 2,
      "stylize": 150
    }
  }
}