Midjourney Features 2026: V8.1 Guide, Pricing & Full Version History

Quick Answer: What's Current in Midjourney Right Now?

Midjourney's Full Version History: V6 to V8.1

If you're arriving at this article because you last used Midjourney around the V6 era, you've missed two major model transitions. Here's the complete, dated timeline so you know exactly where things stand and what changed at each step.

Midjourney V8.1: Every Feature Explained

Here's what actually changed under the hood with V8.1, beyond the headline speed number.

Omni Reference: Character & Style Consistency Explained

Omni Reference is arguably the single most useful feature to survive from V7 into V8.1. It replaced the old --cref character reference flag, which was powerful but genuinely unpredictable — sometimes the character carried through a new generation, sometimes only the style stuck, sometimes it ignored the reference almost entirely.

Omni Reference fixes that by giving you a dedicated reference tab with an explicit strength control, so you tell Midjourney exactly which character or subject to preserve and how strongly the reference should influence the new image.

# Attach a reference image and set its influence strength
a photorealistic portrait of the same person hiking
in the mountains at golden hour, cinematic lighting
--oref [your-reference-image-URL] --ow 350 --cw 100

# --oref  : the reference image (drag into the Omni Reference tab, or link)
# --ow    : reference strength, typically 300–500
# --cw    : character weight — how much of the character (vs style) carries over

It's worth being honest about where this still falls short: Omni Reference holds up well for one consistent character across three or four frames, but tends to drift on longer sequences or full-body shots with multiple characters in frame. If your project needs five or more perfectly consistent reference images, Nano Banana Pro currently handles that specific case more reliably — more on that comparison below.

Draft Mode & Speed Modes: Fast, Relax, and Turbo

Midjourney gives you three speed tiers, and understanding the trade-offs saves real money on your monthly GPU allocation.

• Draft Mode — introduced in V7, carried into V8.1. Roughly 10x faster and half the cost of a standard generation, with voice-command support for rapid ideation. Use it to explore composition and concept before committing GPU time to a full-quality render.
• Fast Mode — the default. Draws from your plan's monthly Fast GPU hour allocation and generates in well under a minute.
• Relax Mode — unlimited but queued, with wait times from a few minutes up to 30 depending on server load. Available on Standard, Pro, and Mega plans for images (Pro and Mega only for SD video). Heavy daily use gradually increases your queue priority wait, though this resets monthly.
• Turbo Mode — up to 4x faster than Fast, but consumes 2x the GPU time per image. Currently not supported on V8.1 Alpha and falls back to Fast automatically when Turbo capacity isn't available.

Midjourney Pricing 2026: Every Plan Explained

Midjourney bills by Fast GPU time, not by an image count — the single most misunderstood part of its pricing. There's still no free trial; it was discontinued in 2023 after abuse. The cheapest way in remains the Basic plan at $10/month.

What each tier actually unlocks

• Basic ($10/mo): Entry point for hobbyists and casual testing. No Relax Mode means once your 3.3 Fast hours are gone, you're waiting for next month's reset — no unlimited fallback.
• Standard ($30/mo): The tier most regular users land on, because it's the first with unlimited Relax Mode for images. Once Fast hours run out, you keep generating — just queued rather than instant.
• Pro ($60/mo): Adds Stealth Mode, which keeps your generations private and out of Midjourney's public Explore gallery. Required by Midjourney's terms for any company earning over $1M in annual gross revenue that wants to use Midjourney commercially.
• Mega ($120/mo): Same feature set as Pro with double the Fast GPU hours — built for studios and agencies running high client volume, particularly if video generation is part of the workflow.

Midjourney V8.1 vs Nano Banana Pro: Which Should You Use?

Google's Gemini 3 Pro Image model — nicknamed Nano Banana Pro — has become the most credible rival to Midjourney's dominance in 2026. They're built for genuinely different priorities, so "which is better" depends entirely on what you're making.

The honest read: choose Midjourney V8.1 for editorial mood, rich colour science, and stylised or cinematic imagery where aesthetic judgement matters more than literal accuracy. Choose Nano Banana Pro for anything requiring exact text, precise editing of existing photos, higher native resolution, or programmatic access via API. Many working creators use both — Midjourney for hero visuals, Nano Banana Pro for anything needing readable text or exact reproduction. If you want a deeper look at how Google's image models fit into a broader AI photo workflow, our Gemini AI photo tutorial covers Nano Banana's editing and generation features step by step.

Prompt Tips That Work in V8.1

Good news if you learned prompting on V6 or V7: nearly everything transfers directly. V8.1 rewards natural, descriptive sentences over disconnected keyword strings, and reads earlier terms in your prompt with more weight — so lead with what matters most.

# Lead with the most important element — V8.1 weights early terms more heavily
editorial product photograph of a ceramic coffee cup on
a marble countertop, soft morning window light, shallow
depth of field, 85mm lens, muted earth tones
--ar 4:5 --stylize 250 --raw

# --ar       : aspect ratio (4:5 for portrait/social, 16:9 for wide)
# --stylize  : 0–1000, how much Midjourney's aesthetic bias applies
# --raw      : strips default styling for literal interpretation

• Write full sentences, not keyword lists. V8.1's natural language understanding means "a woman walking through a sunlit market at dawn" outperforms "woman, market, dawn, sunlight, walking" for coherence.
• Put your priority term first. V8.1 weights earlier words in a prompt more heavily than later ones — lead with the subject or effect that matters most to you.
• Use quotation marks for exact text. "Legible text, clear typography" plus quoted short strings (1–3 words) gives the most reliable text rendering V8.1 offers.
• Add anatomical guardrails for hands. "Anatomically correct hands, five fingers" plus --no extra fingers, deformed hands further reduces the already-improved error rate.
• Use --chaos for exploration, not production. Higher chaos values (0–100) increase variation between your four outputs — useful early, but counterproductive once you know what you want.

Limitations Worth Knowing Before You Subscribe

Who Should Actually Use Midjourney in 2026?

Midjourney V8.1 is worth the subscription if aesthetic quality is your priority and you can justify a recurring cost: content creators, marketers, brand designers, and anyone producing editorial or cinematic visuals where "looks striking" matters more than "matches the brief exactly." It's a weaker fit if you need occasional, low-volume images (a per-image API is cheaper), exact text rendering, or programmatic integration into an app.

For readers building out an AI-assisted content workflow more broadly, our generative AI breakthroughs guide covers how image generation fits alongside text and video tools in a modern creative stack, and our AI comparison tool lets you weigh Midjourney against dozens of alternatives side by side on the criteria that matter to your specific project.

Frequently Asked Questions

Conclusion: V8.1 Is Where You Should Be

If there's one thing to take from this guide, it's this: don't keep building around V6. The gap between V6 and V8.1 isn't cosmetic — it's roughly 40% better anatomy accuracy, native 2K instead of 1024px, 4–5x faster rendering, and a genuinely usable character consistency system in Omni Reference that didn't exist when V6 launched. Every prompting habit you built on V6 or V7 still works; you're just getting materially better output for the same subscription cost.

The practical move: confirm your account is on V8.1 in settings, spend an hour re-running a few of your old go-to prompts to see the difference firsthand, and set up Omni Reference if character or brand consistency matters to your work. From there, the version question mostly answers itself — you'll see why V8.1 became default within the first few generations.

For a wider view of how AI image tools fit into a complete creative or marketing stack in 2026, our generative AI in 2025 guide and Gemini AI photo tutorial are natural next reads.