Midjourney v8.2 Preview AI Image Generation

Midjourney v8.2 Preview is an early-access release of Midjourney's next model, available on FlyAIgh without Discord or a separate Midjourney subscription. Against v8.1 it aims at three things: material lighting that reads more like real light on real surfaces, cleaner local detail in skin, fabric and texture, and characters that stay more recognisably the same across a set of images. Because it is a preview, its behaviour may still change as Midjourney updates it; v8.1 remains available on its own page for production work you want to keep stable.

You can generate from a text prompt or add one reference image alongside it — FlyAIgh sends that image through Midjourney's own image-reference field, so it guides composition, colour and subject the way a reference does in Midjourney itself. Prompts are limited to 1,024 characters, which favours Midjourney's usual short, evocative style over long instructions.

Seven aspect ratios are available: 1:1, 16:9, 9:16, 4:3, 3:4, 3:2 and 2:3. Midjourney has no conventional 1K / 2K / 4K picker; instead FlyAIgh maps the v8.x controls to two tiers. Draft costs 6 credits per image and is the quick way to explore a direction. Fast, the default, costs 8 credits and uses the higher-detail render. The exact cost is shown before you generate.

Midjourney is the model to pick when the aesthetic itself is the deliverable — mood, composition, colour and atmosphere for concept art, editorial illustration, album covers and campaign key visuals. It is weaker than GPT Image 2.6 for exact in-image text and weaker than Nano Banana 2.1 for strict photographic realism.

Try v8.2 Preview when you want the newest look and can live with a model that may still shift; stay on v8.1 for consistency across a project, or v7 for its established style and Turbo tier. All three have their own pages and run on the same FlyAIgh credit wallet.