Best Base Model for LoRA Training in 2026: I Trained the Same Two Faces on Six Models — MesmerTools

Update · July 2026<br>Krea 2 is now #2. After switching its sampling from Krea 2 Raw to Krea 2 Turbo, the same trained LoRA produced much better images, enough to move it from the bottom third all the way to second place. The new results are in the v4 tab of the Krea 2 section below.

A few days ago I got curious about which base model is actually the best at training LoRAs. I run a site that trains custom LoRAs (think AI headshots), it has been on Flux.1 Dev for a long time, and I wanted to know if the newer models were worth switching to. So I let my basement GPUs run for four days straight and trained the same two faces on six different models.<br>Quick Take<br>If you train character or headshot LoRAs and you have been living on Flux.1 Dev, Ideogram 4 is worth a serious look. It held the likeness on both an easy face and a hard one better than anything else I tested, and it is the model I am moving my own pipeline to. Flux.1 Dev is still a perfectly good default. Flux.2 Dev is the one to skip unless you have a big-VRAM card to burn. You can also try trained-model headshots on bestphotoAI, which is the site this whole experiment was for.

Why I bothered<br>Most of these models have been out for a while now, so it felt like a good time to compare them. The thing that pushed me over the edge was CivitAI. If you go looking for character LoRAs on the newer models, people just are not sharing them the way they did during the Flux.1 Dev days. Nobody had really done the boring, apples-to-apples version of this test.<br>I have a couple of RTX 4070 Ti SUPER cards (16GB each) sitting in my basement. With AI doing the driving now, the whole hassle of training is mostly gone: babysitting runs, fixing CUDA errors, tuning rank and learning rate and text-encoder learning rate. So I handed the entire pipeline to Claude, start to finish, and told it to fix its own problems when they came up. One card kept serving a production image model the whole time. The other one did the training.<br>The two subjects<br>First subject is a generic white woman, because most models have no trouble with that. The second is a South Asian man, and this is where it gets interesting. In my experience South Asian and Black faces get much worse resemblance, and the models love to overtrain and slide into racist caricature. You will see exactly that in a few of the runs below, usually at the higher step counts. Same reference photos went into every single model.<br>The two subjects<br>Same reference sets fed to every model. One easy face, one the models historically struggle with.<br>Woman

+8more

Man

+12more

How to read this (v1, v2, v3)<br>Fair warning on the versions. I only noticed after a lot of trainings were already done that my v1 and v2 sample prompts were too basic. A basic prompt has a nasty failure mode: if a model is overtrained, it just hands you back your input photos and everything looks great. That is a lie detector you want pointed the other way.<br>So v3 is the comparison that matters. The v3 prompts are more complex and put the person in novel scenarios (riding a bike through a vineyard, a neon Tokyo street, hiking an alpine ridge). That is how you find out if a model actually learned the face and can recreate it somewhere new, versus just memorizing a handful of pictures. I left v1 and v2 in so you can see the difference yourself.<br>Each model gets its own section below. Toggle the subject (woman or man) and the version, then use the step tabs to walk the checkpoints, newest first. The prompts are seed-locked, so every step renders the same scenes and you can watch the face lock in (and then, often, overcook). Hit "Config & prompts" on any model to see the exact rank, learning rate, and captions it trained on. Almost everything here fit in 16GB, which honestly surprised me. Do not ask me how, Claude figured out the quantization. Flux.2 Dev was the one exception, and it has its own sad story further down.<br>#1Ideogram 4<br>ideogram4 · fp8

Config & prompts

Once I fed it its native JSON caption format, Ideogram 4 held the likeness on both people and stayed convincing on the harder, novel prompts. This is the one I'm moving my pipeline to.<br>WomanMan<br>v3v2v1

RTX 4070 Ti SUPER 16GB2,729 / 3,000 steps4:36:335.47s / it90s / preview

3,0002,7502,5002,2502,0001,7501,5001,2501,000750500250Base

Step 3,000 checkpoint · seed-locked prompts, so every step renders the same scenes.

laughing while riding a bicycle through a sunny vineyard<br>with a serious expression wearing a wool overcoat<br>with a surprised expression on a neon-lit rainy Tokyo<br>with a neutral expression, lit by a soft key<br>focused while cooking at a stove in a bright<br>calmly reading a book in a cozy library<br>smiling while hiking a rocky alpine ridge<br>with a contemplative expression, lit with dramatic Rembrandt lighting

#2Krea 2<br>krea2 · Krea 2 Turbo sampling

Config & prompts

This one flipped on me. Sampled with Krea 2...