Mastering Midjourney: A Complete Guide to Parameters for Stunning AI Art

midjourney blog post image

Understanding Core Midjourney Parameters

Parameters are commands you add to your text prompts to override Midjourney’s default settings. They start with two dashes (–) followed by the parameter name and its value. For example, to create a 3:4 image, you would use –aspect 3:4. Remember to place parameters at the end of your prompt and separate them with spaces.

Version: Choosing the Right Midjourney Model

The version parameter controls which Midjourney model generates your images. Each model has a unique look and understanding of prompts. You can specify the model version using –version or –v followed by the model number. For instance, –v 5 will use the V5 model. Experimenting with different versions can be eye-opening, showing how Midjourney has evolved.

Current values include:

  • –v 3
  • –v 4
  • –v 5
  • –v 5.1
  • –v 6
  • –v 6.1

There are also test models like –test and –testp, and anime-trained models using –ni followed by the model number.

Aspect Ratio: Shaping Your Image

Chaos: Adding Diversity to Your Results

Chaos affects the variety of images in your grid. The parameter is –chaos or –c, with a default value of 0. It accepts whole numbers up to 100, where higher values mean more diverse results. A little chaos, like –chaos 5, can add unexpected twists to your images.

Style and Aesthetics: Fine-Tuning Your Art

Midjourney offers several parameters to control the style and aesthetics of your generated images.

Stylize: Controlling Midjourney’s Aesthetic Influence

Each Midjourney model has a default visual style. The stylize parameter, –stylize or –s, controls the strength of this aesthetic. The default value is 100, with a range from 0 to 1000. Higher values produce more stylized results but may reduce prompt accuracy. Stylize also works with personalization features.

Style Reference: Borrowing Aesthetics from Other Images

Style reference, or –sref, is a powerful feature that lets you apply the aesthetic elements of other images to your creations. You can use image URLs, sref codes, or the word “random” after the parameter. Style weight (–sw) controls the influence of the style reference, with a default of 100 and a range of 0 to 1000. Style version (–sv) is relevant for V6 models when using image URLs as style references.

Style: Raw vs. Stylized

The style parameter, often set as “mode” in default settings, lets you minimize Midjourney’s stylistic influence. Using –style raw can be helpful for more photorealistic images. Older models also have different style codes that affect aesthetics.

Controlling Image Generation Speed and Quality

Midjourney provides parameters to manage the speed and quality of image generation.

Quality: Balancing Detail and Speed

Speed: Turbo, Fast, and Relax Modes

You can control the generation speed with parameters like –turbo, –fast, and –relax. Turbo and fast modes use your fast GPU hours, while relax mode does not (but is not available with all subscriptions). You can set your default speed in settings but override it with these parameters.

Repeat: Running Prompts Multiple Times

The repeat parameter, –repeat or –r, runs the same prompt multiple times. It works with fast or turbo mode. For example, –r 5 runs the prompt five times.

Advanced Parameters for Specific Effects

Midjourney has advanced parameters for specialized effects.

Video: Creating Time-Lapse Generation

The video parameter, –video, creates a short time-lapse of your initial image grid being generated. After the grid is complete, react to it with the envelope emoji to receive a direct message with a link to the video.

Stop: Ending Image Generation Early

The stop parameter, –stop, ends image generation early, taking values between 10 and 100. This can create softer, more abstract results.

Image Weight: Influencing with Images

Image weight, or –iw, is used with image prompts to control how much the images influence the result. The default is 1, and it takes decimal values between 0 and 3 in Midjourney V6.

Character Reference: Matching Characters from Images

Character reference, –cref, matches characters from reference images. After adding the image, the character weight parameter, –cw, controls how closely Midjourney adheres to the character’s clothing and hairstyle.

Personalize: Creating Your Custom Midjourney Model

Personalization, –p, lets you create your own custom Midjourney model based on your style. You first need to teach Midjourney your style, then enable personalization to apply it to your results.

No: Excluding Elements from Your Image

The no parameter, –no, tells Midjourney what to exclude from your image. Separate multiple items with commas.

Tile: Creating Repeating Patterns

The tile parameter, –tile, creates seamless repeating patterns. It doesn’t require any values.

Seed: Controlling the Initial Noise Pattern

Each Midjourney job starts with a visual noise pattern linked to a seed number. The seed parameter, –seed, lets you specify a seed number for consistent results when testing parameter values.

Weird: Injecting the Unexpected

The weird parameter, –weird or –w, creates strange and unexpected results. It takes values up to 3000, with higher values producing more unusual outcomes.

Conclusion

🖥

Do This Locally with Logen

Skip the subscription. Generate images like these on your own GPU — private, unlimited, no cloud dependency.

Try Logen Free →

Midjourney vs Logen

FeatureMidjourneyLogen
Cost$10-60/moFree & Pro plans
PrivacyCloud (images stored)100% local
GenerationsLimited by planUnlimited
Internet RequiredAlwaysNo
Custom ModelsNoYes (any ComfyUI model)
GPU RequiredNo (cloud)Yes (local)
Ease of UseDiscord botDesktop app
Try Logen Free →
FREE 11-PAGE PDF HANDBOOK + LOCAL AI FIELD NOTES

Get the Free 11-Page Local AI Image Generation Handbook

See what GPU VRAM you need for FLUX.1-schnell and SD 1.5, how checkpoints and LoRAs work, and five repeatable workflows on Windows and Linux x64. Enter your email to open the 11-page PDF right away and get our six-part weekly local AI field notes.

6–24 GB VRAM Tables Cloud vs. Local Math 5 Studio Workflows
How signup uses your email

We register your email with the TitanXT newsletter on beehiiv to send the 11-page PDF link, our six-part weekly local AI sequence, and occasional Logen updates. You can unsubscribe in one click from any email or by writing to support@titanxt.io. Read beehiiv's privacy policy (new tab).

Scroll to Top