MODEL Midjourney Last updated:

Midjourney V7

Aesthetic Frontier

Midjourney's seventh-generation image model — released April 2025 on Discord and the new Midjourney web app. V7 still leads on artistic aesthetic against any other text-to-image model: people use it when they want a result that looks designed, not just correct. V8 Alpha entered testing March 2026.

Why it matters

Defines the "good-looking AI image" baseline that every other model is judged against. Even with Flux.2 and Gemini Image catching up technically, Midjourney leads on the subjective artistic-output dimension that drives consumer use.

Core Capabilities

Generative
Produces images, video, audio, or other media.
Multimodal
Combines text, vision, and audio in one model.
Vision
Understands images, scenes, and visual context.

Context Window

Context window not disclosed.

Availability

API
Not available
Product / App
Available
Open Source
Not released
Enterprise
Contact sales

Pricing Model

Subscription
Bundled inside the host product.
Subscription

What it feels like

  • Complete rebuild from scratch — not an incremental update — released April 3, 2025; default June 17
  • Anatomical accuracy jumped 40% (hands, faces) — Midjourney's longest-running weakness finally meaningfully addressed
  • Prompt understanding 35% better — simpler prompts now produce closer-to-intent results
  • Text rendering remains the Achilles heel — only ~15% better than V6, still unreliable for in-image typography
  • Decrypt called it 'the former industry standard struggles to keep up' — improvements are real but rivals caught up
  • Aesthetic quality and personalised style still unmatched for stylised art and illustration use cases

Best use cases

  • Stylised art, illustration, concept art where aesthetics dominate
  • Personalised image style via the personalisation feature (one of MJ's signature differentiators)
  • Mood / vibe-driven creative work where prompt nuance matters more than precision
  • Brand visual exploration and reference generation

Tools to try

Not ideal for

  • Text-in-image rendering (typography, signage, posters with reliable letters)
  • Photoreal product / scene editing requiring precise reference adherence (FLUX.2 / Imagen 4 lead here)
  • API-only headless workflows — Midjourney is interactive-first via Discord and web app