Tests of difrent models i downloaded part2

Decided to do a test run of all checkpoints i have on my disc.
Rules of test:
1. only 1 attempt.
2. using same prompts for all of the pictures.
3. using -1 seed. unles i compared versions of same checkpoint.
4. using same render settings.
5. no lora, only some negative embeds.
6. Vae was added only if the checkpoint recomended one and if it was up to the user i used vae ft mse 840000 ema pruned.
positive:
[well defined female,fox ears,fox tail,sitting on a bench,thighhighs,shorts,hoodie with design,scarf,long messy shiny dark blue hair,long messy twintails,deep in the forest,autumn,windy,lanterns,night sky,night,]
negative:
[(Bybadartist-neg:1.2),(DeepNegativeV1.x easynegative:1.2),(badhandv4:1.2),(low quality, worst quality:1.2),blurry,detached ears,detached tail,human ear,elf ears,demon ears,]
settings:
[Sampling: DPM adaptive steps:80 original resolution: [960x540] CFG: 7 Seed: -1 Hires fix[upscaler: R-ESRGAN 4x+ Anime6B steps 50 denoising 0.3]]
clip skip:2
notes:
-since i use unusual ratio it might show which checkpoints are msot flexible [and not add extra limbs or morph them too weirdly] the ones that did from what i noticed should still do fine if you simply flip the width with height.