Which Let's Enhance model should you use? ✨

Not all images are built the same — and neither are our models. Picking the wrong one is literally the #1 reason people get results they don't vibe with. Let's fix that. 👇

The Quick Answer (for those who won't read the whole thing)

Now let's get into it. 🧵

🌟 Prime — The Default King

Best for: portraits, product photography, nature, fashion, real estate, skin, fur, fabric — basically most photos
Prime is the main character. It's selected by default for a reason — it reconstructs lifelike texture and detail without over-processing or making things look plastic. No sliders, no fiddling. It just works.
It reads what's actually in the image and rebuilds it faithfully — skin pores stay real, fabric grain stays natural, nothing gets invented that wasn't there. If you're not sure what to pick, Prime is your answer 99% of the time.
✅ Use it when: you have a real photo and want the best quality upscale with natural results
❌ Skip it when: your image has flat colors, text, or is AI-generated (it'll add texture where there shouldn't be any)
Pro tip 💡 — if Prime is making skin look too smooth or plastic, lower the Light AI intensity slider to around 10–20%. Game changer.

⚖️ Balanced — The Everyday Hero

Best for: everyday photos that are blurry, compressed, low quality, or generally a bit rough
Balanced is built to significantly improve images that had a tough life — heavy JPEG compression, low megapixels, blur, noise. It reconstructs more aggressively than Prime without going full Ultra. It also handles faces really well, rebuilding features without making people look like mannequins.
✅ Use it when: your photo is clearly low quality and you need visible improvement — travel snapshots, old phone pics, screenshots of photos
❌ Skip it when: your source is already decent — Balanced might over-process it

🪶 Gentle — Less is More

Best for: product photos with text or labels, UI screenshots, posters, menus, infographics, logos, images where exact fidelity matters more than enhancement
Gentle is the most conservative model in the lineup. It adds pixels without interpreting what should be there — geometry stays exact, text stays readable, colors stay true. It won't make your image dramatically "better looking" but it will make it bigger without breaking anything.
✅ Use it when: your image already looks good and you just need more pixels, OR it contains text/logos that absolutely cannot be distorted
❌ Skip it when: your image is low quality and genuinely needs improvement — Gentle won't save it

💪 Strong — For the Rough Ones

Best for: severely blurry, degraded, or low resolution inputs where you need the AI to do heavy lifting
Strong pushes reconstruction harder than Prime or Balanced. It's designed for images that are in rough shape — motion blur, heavy noise, serious compression artifacts. It'll work hard to recover as much detail as possible from not much to start with.
✅ Use it when: your image is genuinely in bad shape and you need the AI to dig deep
❌ Skip it when: your image is decent — Strong can over-reconstruct and introduce artifacts on already good sources

🚀 Ultra — Maximum Power Mode

Best for: professional photography, hero banners, high-end AI renders, catalog images, billboard-scale outputs
Ultra is the most powerful model we have. It reconstructs detail aggressively and comes with two tunable settings — Intensity (how creative the AI gets) and Size of Changes (how broadly it applies those changes). This gives you control over how far it goes.
Ultra is the one that can genuinely surprise you — in a good way on badly degraded sources, and in a bad way on already good images if Intensity is too high.
✅ Use it when: you need maximum quality output and you're not afraid to experiment with the sliders
❌ Skip it when: your image is already sharp — Ultra will over-process it. Also keep Intensity low (10–30%) to avoid the plastic/artificial look
Credit note 💸 — Ultra uses a tiered credit system based on output size. The bigger the output, the more credits it costs. Check the Enhance button before you commit!

🎨 Digital Art — For the Non-Photo Girlies

Best for: illustrations, comics, anime, AI-generated art (Midjourney, SDXL, etc.), flat design, icon sets, vector-style artwork, UI elements
Digital Art was trained specifically on non-photographic content. It understands that flat orange shapes are SUPPOSED to stay flat and orange — not pick up grain or texture. Clean lines stay clean. Solid fills stay solid. Nothing looks accidentally photographic.
It has a Strength slider — set it low when your source is clean and you just need more pixels. Crank it up only if edges are soft and you need more reconstruction work done.
✅ Use it when: anything that's illustrated, drawn, generated by AI, or has flat colors and clean outlines
❌ Skip it when: your image is a real photograph — it'll make it look weird and painterly

📷 Old Photo — Time Machine Mode

Best for: scanned old photos, vintage prints, damaged or faded photographs, B&W images you want restored or colorised
Old Photo is built to repair and restore aged photographs — scratches, tears, fading, discolouration — and can even colorise B&W images automatically. It's the go-to for anything that's been sitting in a shoebox for decades.
Important thing to know ⚠️ — Old Photo has a fixed maximum output of ~1MP regardless of input size. This is intentional! The model prioritises restoration quality over output dimensions.
The recommended two-step workflow:

1. Run Old Photo first to restore and repair ✅
2. Download, then re-upload and apply Prime at 4x to get your resolution back up ✅

✅ Use it when: you're working with genuinely old, faded, damaged, or B&W photographs
❌ Skip it when: your image is modern and in decent shape — it'll reduce your resolution

🧪 Try All — Can't Decide? Same.

Not sure which model is right? Hit Try All and let the platform run your image through multiple models so you can compare results side by side. It costs more credits but it's the fastest way to figure out what actually works best for your specific image. No guesswork, just results. 🙌

The Golden Rules 📌

• 🔍 Always zoom to 100% after processing before downloading — that's where you'll spot any issues
• 📁 Upload PNG when possible — JPEG compression gives the AI confusing artifacts to work with
• 💡 Test on a small crop first — especially when experimenting with Ultra or Strong settings
• 👀 Check the credit cost on the Enhance button before every job — it updates in real time

Still lost? I got you 💬

Hey, I'm Ana from the Let's Enhance Customer Care team 👋
If you've read all of this and you're still not sure which model is right for your specific image — honestly just reach out. I'm here for it. No judgment, no robotic copy-paste replies, just real help from a real human who has seen every possible image situation at this point 😅

Hit me up here:
📧 Email: support@letsenhance.io
Drop your question, tell me what you're trying to do, and we'll figure it out together. We don't gatekeep knowledge around here

— Ana, Customer Care @ Let's Enhance

Updated on: 14/09/2026