AIREITER
Midjourney V8.2 AI Image Generator and API
Midjourney V8.2 is the latest Midjourney image model. Every request returns four separate, full-resolution images, and you can steer them with image prompts, a style reference, or an image to edit, all from one form or one API call.
Four images per request
Each generation returns four independent images instead of a single grid, so you can download or reuse any of them directly.
Standard or native HD
Standard output is 1024×1024 at 1:1 and HD output is 2048×2048. Other aspect ratios scale to the same detail level.
Three ways to use your own images
Add image prompts to guide content and composition, a style reference to borrow a look, or an image to edit with a written instruction.
No parameter syntax to memorize
Pick the aspect ratio, quality, and weights in the form. The Midjourney parameters are added to your prompt automatically.
What to Create with Midjourney V8.2
Illustration and concept art
Explore characters, scenes, and environments with four variations per prompt, then pick the strongest direction.
Consistent brand visuals
Apply one style reference across a whole campaign so every image shares the same colors, medium, and lighting.
Restyling existing images
Turn a logo, sketch, or photo into a pencil drawing, watercolor, or 3D render by describing the change in plain words.
Wallpapers, covers, and social posts
Generate in 16:9, 9:16, 4:3, 3:4, 3:2, or 2:3 so the image fits its destination without cropping.
Supported Midjourney V8.2 Controls
The playground and the API expose the same parameters.
Prompt
Describe the image in natural language. Version, aspect ratio, and quality come from the options, so any -- parameters typed into the prompt are ignored.
Image prompts and image weight
Optional, up to 4 images that guide content, composition, and colors. Image weight (--iw) ranges from 0 to 3, default 1; higher values follow the images more closely.
Style reference and style weight
Optional, one image whose look, colors, and texture are borrowed without copying its people or objects (--sref). Style weight (--sw) ranges from 0 to 1000, default 100.
Edit an image
Optional, one image to change with your prompt as the instruction (--edit). It can be combined with image prompts and a style reference.
Aspect ratio
1:1, 16:9, 9:16, 4:3, 3:4, 3:2, or 2:3. The default is 1:1.
Quality
Standard (--sd) or native HD (--hd). At 1:1 each image is 1024×1024 or 2048×2048.
Midjourney V8.2 Pricing
Per request, not per image
One request produces four images and is billed once. The exact credit cost is shown in the pricing table above and on the Generate button.
Priced by speed
Fast and Relax have different prices. Quality (Standard or HD), aspect ratio, and reference images do not change the price.
How credits are computed
The submit endpoint returns the estimated credit cost before the task runs, and the query endpoint returns the final credits_used after completion.
Failed tasks are free
If a generation fails, no credits are charged.
Generate with Midjourney V8.2
Four steps from idea to four finished images, in the browser or through the API.
Describe the image
Write what you want to see in plain words. For edits, write the instruction, such as "render this logo as a hand drawn pencil sketch".
Add images (optional)
Upload up to 4 image prompts, turn on Style Reference to add a style image, or turn on Edit an Image to add the image to change.
Choose the specs
Pick the aspect ratio and Standard or HD, and adjust image weight or style strength if needed.
Generate and pick
Generation usually takes a few minutes. Download the images you like, or feed one back in as an image prompt to keep iterating.
Use Midjourney V8.2 through the AIReiter API
Submit one normalized task and query the same external task ID until it reaches a terminal state.
Model ID
Send mj_v8_2 as the model value in the request body.
Submit a task
POST /api/openapi/submit with the model, a unique out_task_id, and params: prompt, aspect_ratio, quality (sd or hd), and optionally image_url (up to 4) with image_weight, use_style_ref with style_image_url and style_weight, or use_edit with edit_image_url.
Query the result
POST /api/openapi/query with the same out_task_id until the status is completed or failed. Completed responses include four image URLs and the final credits_used.
Store results on your side
Download the output images and keep them in your own storage instead of hot-linking the result URLs long term.
Related AI Image Models on AIReiter
Midjourney V8.1
The previous Midjourney V8 release. Choose V8.2 for the newest model, image editing with written instructions, and four separate images per request.
Midjourney V7
The earlier Midjourney generation, with retexture and background removal as separate models.
Nano Banana 2.1
A fast image model with accurate in-image text and up to 10 reference images. Choose it for posters and layouts with readable copy.
Midjourney V8.2 FAQ
Answers about capabilities, controls, pricing, and API usage.
/ 01What is Midjourney V8.2?
Midjourney V8.2 is the latest Midjourney image model. On AIReiter it generates four separate images per request from a text prompt, optionally guided by image prompts, a style reference, or an image to edit.
/ 02Do I need to write Midjourney parameters like --ar or --v?
No. Choose the aspect ratio, quality, and weights in the form or as API params, and the parameters are added for you. Parameters typed into the prompt are ignored.
/ 03What is the difference between Standard and HD?
At 1:1, Standard images are 1024×1024 and HD images are 2048×2048. Other aspect ratios scale accordingly, and the API returns each image's real size.
/ 04What is the difference between an image prompt, a style reference, and editing?
Image prompts guide content, composition, and colors. A style reference borrows only the look, such as colors, medium, and texture. Editing changes one existing image according to your written instruction.
/ 05How many reference images can I use?
Up to 4 image prompts, plus one style image and one image to edit per request.
/ 06Which aspect ratios are supported?
1:1, 16:9, 9:16, 4:3, 3:4, 3:2, and 2:3.
/ 07How is Midjourney V8.2 priced?
Once per request, and each request returns four images. Only the speed (Fast or Relax) affects the price, and failed tasks are not charged. See the pricing table on this page for current credits.
/ 08Which API model ID should I use?
Use mj_v8_2 with POST /api/openapi/submit, then poll POST /api/openapi/query with the same out_task_id.