🎨 MidjourneyBeginner

Midjourney V8.1: How to Use Faster Renders, HD Images, and Raw Mode

A practical guide to Midjourney V8.1: choosing SD or HD, using Raw mode for stronger prompt adherence, and avoiding common cost and editing mistakes.
✍️ GoToUseAI📅 Updated 2026-07-228 min read
★★★★½
4.7· 94 readers found this helpful

Midjourney V8.1 became the default version in June 2026. Midjourney describes it as its fastest model so far, with better prompt reading and small-detail retention. The practical question is not whether to switch it on—it is how to choose the right mode before you spend GPU time.

Start With the Right Version and Mode

Open Midjourney settings and confirm that V8.1 is selected. For a first pass, use the standard image size. It is the cheapest way to explore composition, lighting, subject, and style.

Use Raw when prompt adherence matters more than Midjourney’s default aesthetic styling. Raw is useful for product images, editorial compositions, specific camera directions, and concepts where a strong default “look” would get in the way.

Example:

small ceramic coffee cup on a pale blue studio pedestal,
soft side lighting, clean e-commerce product photography,
no text, no hands, generous negative space --raw

Raw does not make a weak prompt precise. It simply gives your instructions more influence.

SD vs. HD: Make the Choice Late

V8.1 offers HD images at 2K resolution without a separate upscale step. Midjourney’s documentation notes that HD costs more GPU time than SD. Use SD while you are still deciding the concept, then rerun the winning prompt in HD.

A cost-aware workflow looks like this:

  1. Generate several SD concepts.
  2. Choose the strongest composition.
  3. Make one targeted change at a time.
  4. Rerun the chosen seed in HD.
  5. Export or prepare the final image.

Do not generate everything in HD “just in case.” Most early images are discarded, so high resolution at the exploration stage usually wastes time.

Write Prompts in Layers

The easiest prompt structure is:

[subject] + [setting] + [visual direction] + [composition] + [constraints]

For example:

friendly robotics engineer in a bright workshop,
editorial documentary photography, natural window light,
medium-wide composition, tools softly out of focus,
blue and warm orange color palette, no logos --raw

Each layer answers a different question. The subject tells Midjourney what matters; the setting and visual direction establish context; composition controls the frame; constraints prevent common mistakes.

Preserve a Good Result With Seeds

When an image is close to what you want, save the prompt and seed before experimenting. A seed lets you return to the same starting point while you change one variable—such as the aspect ratio, a color, or the background.

This matters for campaigns and product sets. Without a saved seed and prompt, each new render becomes a fresh lottery ticket.

Know the HD Editing Limitation

Midjourney documents an important trade-off: using inpainting, outpainting, Pan, Zoom Out, Edit, or Vary Region on an HD V8.1 image can reduce the resulting image to SD. Plan detailed edits before your final HD render where possible. If you need an edit after HD, check the result and upscale again if necessary.

A Simple Product Visual Workflow

  1. Write a concise product brief: audience, setting, mood, and required empty space.
  2. Generate four SD directions with the same subject.
  3. Pick one direction and use Raw mode if the prompt needs tighter control.
  4. Save the best seed and prompt.
  5. Rerun the selected image in HD.
  6. Check text, hands, labels, and factual product details before publishing.

AI images are excellent for concepts and supporting visuals, but they are not a substitute for checking brand claims, trademarks, or product accuracy.

For current model behavior, costs, and version availability, use Midjourney’s official V8.1 documentation. Those details can change faster than a static tutorial.

#Midjourney V8.1#Midjourney HD#Raw mode#AI image generation#prompt writing

Discussion

📚 Continue Learning