Three Image Models, One Studio: Which One to Pick for Which Picture

Open the image studio and, under the prompt box, there is a select labelled Model with three entries in it. The names are not descriptive, and the only hint at the difference is the coin figure under the Generate button changing as you switch between them. This post is the description the select does not have room for.

Every number here is read from the catalogue the server bills with, not typed, so if the figures on the product page ever move, this page moves with them. We are not going to claim one model is faster than another. We know what each costs and what each renders; we do not time them, so we do not sell speed.

The full model

This is the model that has made every picture on Fellowi since the studio opened, and the one the select shows first. Standard quality renders at about 1.5K and costs 40 coins; High renders at about 2K for 60. It accepts reference images, so it is what runs when you attach your own photo and ask for an edit, and the prompt enhancer is available on it. When a picture is going somewhere that matters, a print, a cover, a client, this is the default for a reason.

The mini model

A step down in size and a step down in price. Standard renders at about 1K for 30 coins, High at about 2K for 50 coins from text, or 60 when you attach references and it becomes an edit. Everything else is intact: both quality tiers, both file formats, the prompt enhancer, reference images. If you have never changed the select, you have never seen what it does, and for a social post or a thumbnail the difference from the full model is smaller than the coin difference.

The budget model

The third entry is deliberately limited, and we would rather say so here than let you find out at the Generate button. It costs 1 coins per picture, renders a single small size of about 0.5K, and saves only as JPEG. There is no High tier to switch to, the prompt enhancer is off, and it does not accept reference images at all: it is text-to-image only, so switching to it with a photo attached quietly drops the photo and hides the attach control. It also exists only in the studio; the public API does not offer it.

What it is for is volume. At 1 coins you get 40 pictures for the price of one Standard on the full model and 60 for the price of one High. That is the right trade when you are still working out what you want: trying ten compositions, checking whether a prompt reads the way you think it does, picking a direction before spending real coins on it. The draft-first workflow was written with exactly this model in mind.

One more honest line. The budget model runs on a separate, smaller service, and when that service is overloaded the studio says so and returns your coins rather than quietly making the picture on another model. Try again a little later, or pick the mini or the full model yourself; the choice stays yours.

The surcharges are the same on all three

Two options add coins on top of the model price wherever they are available: the prompt enhancer adds 10 coins, and each attached reference image adds 10. The budget model has neither, which is part of why it stays cheap. The reference image guide explains when the second surcharge is worth paying, and the honest answer is: when the face or the product has to stay the same, and not before.

A plain rule for choosing

• Exploring an idea: the budget model. Many small pictures, few coins, and nothing lost if nine of ten go in the bin.
• Making the picture you will actually use: the mini model at Standard, then the full model at High only for the one that has to be large.
• Editing your own photo or holding a face: the mini or the full model, because they are the ones that take references. The consistent character guide covers the second case.
• Working through the API: the full or the mini model. The budget model is a studio-only shortcut.

What the free coins buy

A new account starts with 60 coins. On the budget model that is 60 pictures, on the mini model 2 at Standard, on the full model 1. Which is to say: the first afternoon in the studio is best spent on the small model, learning how prompts land, and the first paid coins on the bigger ones. The quality tiers and file formats themselves are covered in quality and formats, and the coins in coins explained. Both are short.