Sign In

Royal Blend XL (Anima + Illustrious)

Download

1 variant available

bf16 SafeTensor

Royal_Blend_Anima_V1.safetensors

BF16, good balance (pruned) • 5.24 GB

Verified:

Type
Checkpoint Merge
Stats

409

Reviews
Published

Jul 14, 2026

Base Model

Anima

Hash
AutoV2
21556D7C8F
Trigger Words
anime screenshot, anime coloring
default creator card background decoration
Followers - 1799

1.8K

Likes - 13889

13.9K

CivChan Stabby Stabby Valentine 2026

License:

Anima

The Anima Model is licensed by CircleStone Labs LLC. Copyright CircleStone Labs LLC. IN NO EVENT SHALL CIRCLESTONE LABS LLC BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH USE OF THIS MODEL.

Built on NVIDIA Cosmos

Update

Anima V1.0

This is the first release based on Anima and features a clean 2D anime style. Like previous IL based version, the main focus of this model is to create girls with "big" proportions.
Adding tags like wide hips, thick thighs or similar will further emphasize the proportions.

The model has text encoder and VAE embedded, so no need to load external. Please note that I only tested this in Forge Neo. If it does not work on Comfy UI, please load them externally as usual.

A turbo lora is also baked in for faster image generation.

Recommeded settings:

Sample/ Scheduler: Euler A + normal

Steps: min. 20 for normal generation / 8-12 for turbo generation
CFG: 3-5 for normal generation / 1-2 for turbo generation

For best results, please add the tags anime screenshot, anime coloring to the positive prompt.

More complex compositions might require you to reduce the CFG to 1.5-2.0 in order to get desired results when using normal generation step count.

Version 6.0

In this release, the style shifted slightly to a bit more semi-realistic look. The body proportions can get now far more thicker and voluptuous, with either fat, muscles or both combined. And the breasts can of course also get quite large.


Tested with DPM++ 2M SDE and Karras

Version 5.0
This version makes small changes to the style and let it look a bit more saturated and a bit more "smoother".
But most changed were made in how the breasts are displayed. They now have a more natural look and can get "pretty big"!

Version 4.0
For this release, I took Royal Blend V2 as the base and added Hotaru Blend V2 and some styling loras to the merge. The style is now far more detailed and less "cutesy" then V3 was.

I recommend using a Highres Fix for this release, as it improves the finer details drastically.

The following settings were used for the Sample pictures using Forge:
Sampler / Scheduler: DPM++ 2M with Karras or Euler A+SGM Uniform
CFG: 4-5 / 25 steps
Highres fix: 1.5 scale, 15 highres steps, 0.5 denoise, Upscaler: Latent (nearest-exact)

Version 3.0

This new version is a Rouwei/Illustrious based merge.

It features

  • a more colorful and cleaner look

  • better poses and anatomy

  • improved facial details

  • improved background details

New recommended settings

Sampler: Euler A / Scheduler: SGM Uniform

Steps: 25-30

CFG: 3-5

Resolutions: SDXL recommended resolutions (832x1216, 896x1152...)

Resolutions based on 768x768 work also well, recommended for slower machines

Quality tags:

Positive:

masterpiece,best quality,amazing quality

Negative:

low quality, worst quality, watermark, signature, censored, mosaic censoring, bar censor

Version 2.0:

This Version is a completly new merge with the style now shifted to a bit more mature and sexy look of the characters. But it still kept a bit of the cute style.

  • more mature look of the characters

  • more detailed backgrounds

  • colors are more saturated

  • increased skin details

Version 1.0

This IL/Noob-based checkpoint is a personal merge from me. It combines serveral models that I like and is intended to produce a cute, kawaii style, increased skin details and detailed backgrounds.

The model is EPS based.

VAE is included

Recommended settings:

Sampler: Euler A / Scheduler: SGM Uniform

Steps: 25-40

CFG: 3-7

Resolutions: SDXL recommended resolutions (832x1216, 896x1152...)

optional: High-res fix and adetailer depending on use case

Quality tags: Smooth Pos. Illustrious+ for positive prompt

Smooth Neg. Illustrious for negative prompt

List of merged models:

RedFuku 2025.1.1

IllusionBreed "Slender" V3.0

coco-Illustrious-XL V6

FenrisXL VAE

merged with Reforge Checkpoint merger

Receipe:

1st step: RedFuku 2025.1.1 + IllusionBreed "Slender" V3.0 with 50% ratio

2ns step: result of step 1 + coco-Illustrious-XL V6 with 50% ratio

Permission: no commercial use, no selling, use only on Civitai (please also check the sources for their respective permissions)