Model comparison

Anima vs stable-diffusion-3.5-medium

At Q4_K_M, Anima is the smaller download — 1,384,407,168 bytes against 1,787,064,768.

From the file· summed bytes, KV per layer

Side by side

Animastable-diffusion-3.5-medium
Parameters2.1B2.5B
Architecturecosmos
Layers
Native context
Mixture of expertsnono
Quantizations published2321
Smallest quantization0.95 GiB1.35 GiB
Q4_K_M1.29 GiB1.66 GiB
Licenceotherother

KV cache by context

the term that decides long-context viability
ContextAnimastable-diffusion-3.5-mediumRatio
4,096
8,192
16,384
32,768
65,536
131,072
Anima vs stable-diffusion-3.5-medium — size, memory and hardware fit — ossmodeldb