Midjourney Aspect Ratio Guide: --ar Syntax, Best Ratios, and When to Use Each

Four rectangle outlines of different proportions side by side, labeled square, portrait, landscape, and tall, representing different Midjourney aspect ratios 1:1 4:5 16:9 9:16 Same prompt, four canvas shapes, four different compositions

What you will learn

  • The syntax is --ar [width]:[height], added at the end of your prompt
  • If you omit --ar entirely, Midjourney defaults to a 1:1 square
  • Midjourney's supported ratio range runs roughly 1:2 to 2:1 - always check docs.midjourney.com for the current exact limits
  • Aspect ratio changes composition and subject placement, not just canvas shape - wide and tall ratios push Midjourney toward genuinely different framing decisions
  • Changing ratio after generation only crops the image - it doesn't regenerate a native composition for the new shape

Quick Answer

The Midjourney --ar parameter sets your image's dimensions using the syntax --ar [width]:[height], for example --ar 16:9. If you leave it off, Midjourney generates a 1:1 square by default. Beyond just picking a canvas shape, --ar genuinely changes how Midjourney composes the scene - wide ratios favor environmental context, tall ratios favor close subject framing, and square forces centered balance. This guide covers the full syntax, the supported ratio range, a use-case comparison table, and exactly how each ratio reshapes your composition.

In this guide

  1. What is the Midjourney --ar parameter
  2. How to change aspect ratio in Midjourney
  3. Midjourney's aspect ratio range and default
  4. Best aspect ratio for Midjourney by use case
  5. How aspect ratio changes composition and subject placement
  6. 16:9 vs 4:5: a quick comparison
  7. Common aspect ratio mistakes

Most people treat the Midjourney --ar parameter like a resize tool - pick a shape, get an image in that shape. That's true, but it's only half the story. Aspect ratio also tells Midjourney how to compose the scene in the first place: what gets cropped in, what gets room to breathe, and where the subject sits in the frame. Get the ratio right before you generate, and you get a native composition built for that shape. Get it wrong and try to fix it after the fact, and you're just cropping away parts of an image that were composed for a different shape entirely.

This guide covers the mechanics - syntax, range, and defaults - plus the part most references skip: how each ratio actually reshapes composition and subject placement.

What is the Midjourney --ar parameter

--ar sets the width-to-height dimensions of your output image. It's one of the most-used parameters in Midjourney, because it applies to essentially every generation you'll ever run - there's no image without some aspect ratio attached to it.

Basic syntax:

[your prompt text] --ar [width]:[height]

Example

'a woman walking through a rainy city street at night, cinematic lighting --ar 16:9'

--ar is one of a much larger set of Midjourney parameters - version, stylize, chaos, raw, seed, and more. If you want the full reference covering every parameter and how they combine, see our complete Midjourney parameters guide. This guide goes deep on --ar specifically, since aspect ratio deserves more room than a quick mention inside a 15-parameter roundup.

How to change aspect ratio in Midjourney

The correct way to change aspect ratio is to set --ar before you generate, not after:

  1. Decide on your final use case first - a square Instagram post, a wide YouTube thumbnail, a tall Story - before you write the prompt.
  2. Write your scene prompt as normal, describing subject, setting, lighting, and mood.
  3. Add --ar [width]:[height] at the end of the prompt, matching the ratio you decided on.
  4. Generate. Midjourney composes the scene natively for that shape - subject placement, negative space, and framing are all built around the ratio you specified.
  5. If you need a different ratio afterward, don't crop - change the --ar value and regenerate instead. Cropping removes real composition, it doesn't rebuild it.

That last point is worth repeating because it's the single most common aspect ratio mistake: treating --ar as something you can fix in post. It isn't. It's a generation-time decision, not an editing-time one.

Midjourney's aspect ratio range and default

If you don't include --ar in your prompt at all, Midjourney defaults to a 1:1 square. That's the fallback shape for every generation unless you specify otherwise.

Midjourney's supported ratio range runs roughly from 1:2 to 2:1 - meaning one side of the image can be up to about twice as long as the other, in either direction (tall or wide). Ratios more extreme than that are typically rejected or automatically adjusted back within range. Because Midjourney periodically adjusts model behavior, treat this as an approximate range rather than a permanently fixed number, and check docs.midjourney.com if you're working with an unusual, non-standard ratio and need the current exact limits.

Custom ratios within that range work fine too - --ar 5:4 or --ar 7:4 are both valid inputs, not just the handful of "standard" ratios listed below. Midjourney accepts any whole-number ratio inside its supported range.

Best aspect ratio for Midjourney by use case

Here's how the common ratios map to real use cases, along with a quick note on how each one tends to shape composition:

Ratio Platform / Use-Case Tag Composition Note
--ar 1:1 Square - Instagram feed, profile images, Amazon listings Forces a centered, balanced composition - the subject usually sits dead-center with symmetric space on all sides
--ar 4:5 Portrait - Instagram feed posts (most common social format) Tighter vertical framing, subject fills more of the frame with less background visible top and bottom
--ar 9:16 Stories, Reels, TikTok, Pinterest Strong vertical emphasis, close subject framing, minimal room for wide environmental context
--ar 16:9 YouTube thumbnails, website banners, widescreen, landscape Wide environmental framing with generous negative space around the subject, favors context over close-up detail
--ar 2:3 Print, portrait photography, editorial Classic portrait photography proportions - moderate vertical emphasis, natural for single-subject portraits
--ar 3:2 Standard landscape photography, website imagery Balanced horizontal framing - wide enough for scene context without the extreme negative space of 16:9

How aspect ratio changes composition and subject placement

This is the part most parameter references skip entirely: --ar isn't just a canvas shape, it's a composition instruction. Midjourney doesn't generate one "true" image and then crop it to fit your ratio - it builds a different composition depending on the shape you give it.

  • Wide ratios (16:9, 3:2) push toward environmental, contextual framing. With more horizontal room to fill, Midjourney tends to add background detail, wider negative space, and a sense of setting - the subject becomes one element within a scene rather than the entire scene.
  • Tall ratios (9:16, 2:3, 4:5) push toward close, vertical subject emphasis. With less horizontal room, there's simply less space for background - the subject gets pulled closer, tighter, and more dominant in the frame, often at the cost of surrounding context.
  • Square (1:1) forces centered, balanced composition. Equal width and height nudge Midjourney toward symmetric placement - subject centered, matching space on left/right and top/bottom - which reads as calm and formal compared to the more dynamic framing wider or taller ratios produce.

Practical example: prompt the same scene - "a woman standing in a garden at golden hour" - with --ar 16:9 versus --ar 9:16. The 16:9 version will likely show more of the garden, more depth, more negative space around her. The 9:16 version will likely crop in tighter on her figure, show less of the garden's width, and emphasize her vertically in the frame. Same words, genuinely different photograph - not just a different-shaped version of the same photograph.

16:9 vs 4:5: a quick comparison

These two ratios come up constantly and get confused often, so here's the direct comparison:

  • --ar 16:9 - use for YouTube thumbnails, website banners, widescreen or cinematic shots, and any image where the setting matters as much as the subject. Composition leans environmental, with room to show context around the subject.
  • --ar 4:5 - use for Instagram feed posts and any shot where the subject should dominate the frame. Composition leans tight and vertical, with the subject filling more of the visible space and less room for background.

If you're unsure which to pick, ask whether the setting or the subject is doing more of the storytelling in this shot. Setting-forward shots want 16:9. Subject-forward shots want 4:5.

Common aspect ratio mistakes

A few --ar mistakes come up over and over:

  • Cropping instead of regenerating. As covered in our parameters guide's pro tip, cropping after generation almost always loses composition elements Midjourney placed intentionally for the original ratio. If you need a different shape, set --ar and generate again rather than cropping the result.
  • Assuming only the "standard" ratios work. Custom ratios like --ar 5:4 or --ar 7:5 are valid as long as they fall within Midjourney's supported range - you're not limited to the handful of common ratios listed in most tutorials.
  • Picking a ratio based on habit rather than the shot. Defaulting to 1:1 or 4:5 out of habit, even when the shot is clearly an environmental, wide scene that would compose better at 16:9 or 3:2, leaves real composition quality on the table.
  • Forgetting --ar entirely. Without it, every generation defaults to a 1:1 square - fine for some use cases, but easy to forget and end up with the wrong shape for your actual destination.

None of this is complicated once it clicks - the core habit is simply deciding on your final use case before you generate, and letting --ar drive the composition from the start rather than trying to fix the shape afterward.

Want help setting the right --ar alongside the rest of your prompt - subject, colors, composition, lighting, and camera? The free Midjourney Prompt Builder includes a visual aspect ratio selector so you can preview the shape before you generate, plus the rest of the 6-layer framework to build a complete prompt around it.

Frequently asked questions about the Midjourney aspect ratio parameter

1:1 (square) is Midjourney's default aspect ratio. If you don't add an --ar parameter to your prompt at all, every generation comes back as a square image.

Midjourney supports roughly a 1:2 to 2:1 ratio range - meaning one side can be up to about twice as long as the other. Extremely elongated ratios beyond that range are typically rejected or automatically adjusted, so always check docs.midjourney.com for the current exact limits before relying on an unusual ratio.

You can't truly change the aspect ratio after generation - cropping is the only after-the-fact option, and cropping removes content rather than regenerating a native composition for the new shape. If you need a different ratio, the correct approach is to set --ar to the ratio you want and generate again, since Midjourney composes the shot differently for each ratio rather than just resizing the canvas.

For an Instagram feed post, --ar 4:5 is the best choice - it's the platform's preferred portrait ratio and takes up the most vertical space in-feed. Use --ar 1:1 for a classic square grid look, or --ar 9:16 if the image is destined for Stories or Reels instead of the main feed.

Use --ar 16:9 for wide, environmental, landscape-style shots where context and setting matter as much as the subject - YouTube thumbnails, banners, cinematic scenes. Use --ar 4:5 for portrait-style shots where the subject should dominate the frame with tighter, closer framing - Instagram feed posts, single-subject portraits, product close-ups.