New users can enjoy 50% off if you subscribe within one hour during our Early Bird deal!
00min
:
00sec
:
00ms

Midjourney v7 vs v6: The Power of Prompting in My Native Language

Published: April 14, 2025
Blog
Midjourney v7 vs v6: The Power of Prompting in My Native Language

Introduction

For the past few days, I’ve been experimenting with Midjourney v7, specifically testing how well it handles prompts in my native language (Bahasa Indonesia). The results? Impressive.

Previous versions of Midjourney struggled with non-English languages, often producing random or disconnected outputs that missed the cultural and linguistic nuances. But v7 is different—it demonstrates a vastly improved understanding of context, culture, and even racial representation based on the language used.

Here’s what I discovered.

1. Vastly Improved NLP for Non-English Languages

One of the most noticeable upgrades in Midjourney v7 is its natural language processing (NLP) capabilities for non-English prompts.

  • In v6.1, prompting in Bahasa Indonesia often led to unpredictable results—sometimes the images were completely off, indicating that the model didn’t truly “understand” the intention behind the words.
  • Now, in v7, the model grasps context much better, suggesting it has been trained on a more extensive multilingual dataset.

This improvement means that language is no longer just a translation layer—it’s a cultural and contextual signal that influences the output.

2. Stronger Cultural & Racial Understanding

Another fascinating aspect of v7 is its ability to interpret cultural and racial context based on the language used in the prompt.

Ethnic Representation: Beyond Default Assumptions

When I prompted “a woman” in Bahasa Indonesia, Midjourney v7 generated a face with Southeast Asian (Malay-Indonesian) features—not East Asian, South Asian, or Caucasian.

This is a huge leap compared to other AI image models like:

  • Reve
  • FLUX (without LoRA)
  • Hailuo image-o1

These models tend to default to Caucasian features unless explicitly instructed otherwise. But v7 adapts to the linguistic context, producing results that align with the cultural background embedded in the language.

(Side note: While I personally prefer East Asian features—my wife is Indonesian-born Chinese—it’s great to see the model accurately representing different ethnicities based on language!)

Cultural Nuances in Abstract Concepts

The cultural adaptation isn’t limited to faces. Even abstract concepts like “wedding dress” change based on language.

  • When I prompted in English, I got a Western-style white wedding gown.
  • When I used Bahasa Indonesia, the output reflected traditional Indonesian wedding attire, complete with cultural elements like kebaya or songket fabrics.

This shows that v7 doesn’t just translate words—it understands the cultural references tied to them.

Creative Prompt Showdown: Midjourney v6 vs v7 🎨

We ran a series of prompts through both v6 and v7 to compare speed, composition, and depth. The results speak for themselves.

A fluffy puppy playing in the snow on a moonlit night in the mountains

Albert Einstein doing a science experiment

Portrait of Marie Curie in her lab

Child looking at an insect through a magnifying glass

A woman at work

Final Thoughts

Midjourney v7 is a game-changer for non-English users.

  • Better NLP means more accurate interpretations of prompts.
  • Cultural & racial awareness ensures outputs align with linguistic context.
  • Nuanced understanding of abstract concepts makes it more versatile for global users.

For creators who work in non-English languages, v7 finally feels like it was built for us, not just English speakers.

Using Midjourney v7 on ImaginePro

Transitioning to v6 is a breeze. While your default may still be v6, you can easily switch to v7. Here’s how:

Temporary Switch: Simply add —-v 7.0 at the end of your prompts to access v7 features for that session.