animagine-xl-4.0-opt_clear

Animagine-XL_4.0_Opt_clear

• Guide (External site): English | Japanese
• Ciivitai
• Animagine-XL_3.0_clear

<div style="display: flex; justify-content: center; align-items: center; gap: 2em;"> <div> <img src="./images/Hitman.png" alt="Hitman" width="400" height="566"> </div> <div> <img src="./images/2_girl_students pixel-art.png" alt="2 girl students" width="400" height="566"> </div> </div>

Compare to Animagine-XL_4.0_Opt (Original)

Animagine-XL_4.0_Opt_clear | Animagine-XL_4.0_Opt (Original)

<div style="display: flex; justify-content: center; align-items: center; gap: 2em;"> <div> <img src="./images/Animagine-XL_4.0_Opt_clear_1.png" alt="Animagine-XL_4.0_Opt_clear_1" width="400" height="400"> </div> <div> <img src="./images/Animagine-XL_4.0_Opt_original_1.png" alt="Animagine-XL_4.0_Opt_original_1" width="400" height="400"> </div> </div> <div style="display: flex; justify-content: center; align-items: center; gap: 2em;"> <div> <img src="./images/Animagine-XL_4.0_Opt_clear_2.png" alt="Animagine-XL_4.0_Opt_clear_2" width="400" height="400"> </div> <div> <img src="./images/Animagine-XL_4.0_Opt_original_2.png" alt="Animagine-XL_4.0_Opt_original_2" width="400" height="400"> </div> </div> <div style="text-align: center; margin-left: auto; margin-right: auto; width: 800px; max-width: 90%;"> <img src="./images/Animagine-XL_4.0_Opt_clear_2_compare_close_up.png" alt="Close-up comparison showing finer details and sharper edges in Animagine-XL_4.0_Opt_clear anime outputs." width="800" height="400"> </div>

A fine-tuned version of Animagine-XL 4.0 Opt model, designed to produce clearer and more vibrant outputs.

Compared to the original, this model generates illustrations with higher contrast, richer colors, and more refined details.

INT8_ConvRot_HQ (light weight)

INT8_ConvRot_HQ is a lightweight version. The CLIP and VAE remain unchanged, while only the intermediate layers of the UNet have been converted to INT8_ConvRot format.

Image quality is slightly lower than the original v2, but it is particularly well-suited for fast generation in low-VRAM environments (4-6 GB).

INT8_ConvRot is currently supported in ComfyUI and Stable Diffusion webUI Forge NEO.

---

Recommended Settings

General

• CFG scale: 4-6
• Sampler: DPM++ 2M SDE (Euler a)
• scheduler: beta
• Steps: 15-25 (around 20)

Quality Enhancement Tags (same with Animagine-XL 4.0)

masterpiece, high score, great score, absurdres

Recommended Negative Prompt (same)

lowres, bad anatomy, bad hands, text, error, missing finger, extra digits, fewer digits, cropped, worst quality, low quality, low score, bad score, average score, signature, watermark, username, blurry

Upscaler

your choice, examples:

• RealESRGAN_x4Plus Anime 6B (Commercial use)
• IllustrationJaNai_V1_ESRGAN (No-Commercial use)

Detailer

works well.

---

Tips (External site)

• Steps and Scheduler: English | Japanese
• Upscalers: English | Japanese
• Negative Prompt: English | Japanese
• Improving Prompt Encoding in SDXL: English | Japanese

---

License

• CreativeML Open RAIL++-M