Finally, a model that actually gets women's ages right: MatureRitual ANIMA

Town Square / 小镇广场 General Chit-Chat / 日常闲聊
1

I was browsing Civitai and came across MatureRitual (熟メス儀式 ANIMA). The author’s opening line really resonated with me: “We need models that truly understand women of different ages, not just ones that can only generate lolis.” The whole point of this model is to draw mature women and actually grasp a diverse range of age groups.

Instead of vague tags like “1girl,” it uses age-specific tags. For women, you’ve got bishoujo, mature lady, lady, etc. For men, it’s bishounen, gentleman—replacing the old 1boy/1girl approach.

One thing you absolutely have to watch out for: use the animaLLMLayerwise CLIP, NOT Qwen3_06B. The author stressed this multiple times. For the positive prompt, skip any quality words. For the negative, throw in score_1 to score_3 plus an artist name. Recommended settings: CFG 5, 30 steps, euler a for better anatomy, dpmpp_2m_sde_gpu might give better prompt adherence, scheduler simple, and resolution at 832×1216 or higher.

I’m totally on board with the idea—models shouldn’t be stuck on just one age group.

2

Yeah, that’s a real issue. Age diversity has always been a weak spot.

4

Using age tags instead of 1girl — that’s actually a pretty clear way to think about it.

5

If you don’t put any quality tags in the positive prompt, that’s pretty standard for the anima style.

6

Two samplers splitting the work - one handles anatomy, the other handles prompt adherence.

7

Starting at 832x1216, vertical framing works best for portraits.