by Midjourney
Midjourney's smartest and most coherent model. Enhanced prompt understanding, voice prompting, draft mode for rapid iteration.
Midjourney v7, released April 2025 and refined through early 2026, is described as their "smartest, most coherent model yet." Major improvements in prompt understanding, anatomy, and a new draft mode for rapid iteration.
V7 understands longer, more descriptive prompts than v6:
A lone astronaut standing on the edge of a massive crater on Mars, looking down at a bioluminescent lake of alien liquid, dramatic scale showing the tiny figure against the vast landscape, volumetric dust, golden hour lighting from twin suns --ar 16:9 --v 7
| Parameter | Effect | Recommended |
|---|---|---|
| --v 7 | Use version 7 | Always specify |
| --ar 16:9 | Aspect ratio | Match your use case |
| --stylize 100-1000 | Artistic interpretation | 250 (balanced), 750 (artistic) |
| --chaos 0-100 | Variation between results | 10-30 for controlled variety |
| --quality 0.25-2 | Render quality/time | 1 (default), 0.5 for drafts |
| --no [item] | Negative prompt | --no text, watermark |
| --sref [url] | Style reference image | Maintain consistent aesthetic |
Generate quick previews at fraction of the cost:
/imagine prompt: your prompt here --quality 0.25
Then upscale the winners to full quality.
Maintain consistent style across multiple images:
/imagine prompt: a cozy mountain cabin --sref https://example.com/reference-style.jpg
| Use Case | Quality |
|---|---|
| Artistic/stylized images | Excellent (best in class) |
| Fantasy and sci-fi | Excellent |
| Portraits and fashion | Excellent |
| Architecture | Very Good |
| Photorealism | Very Good (Flux 2 is more realistic) |
| Text in images | Good (Flux/DALL-E are better) |
Quick tips from the community about what works with Midjourney v7 right now.
Sign in to share a tip.
No tips yet. Add a tip for this model.