Model comparison

Wan_2.1_ComfyUI_repackaged vs Wan2.2-I2V-A14B

At Q4_K_M, Wan2.2-I2V-A14B is the smaller download — 9,651,728,896 bytes against 68,351,109,664.

From the file· summed bytes, KV per layer

Side by side

Wan_2.1_ComfyUI_repackagedWan2.2-I2V-A14B
Parameters14.3B
Architecturepigwan
Layers
Native context
Mixture of expertsnono
Quantizations published4830
Smallest quantization4.25 GiB4.94 GiB
Q4_K_M63.66 GiB8.99 GiB
Licenceapache-2.0apache-2.0

KV cache by context

the term that decides long-context viability
ContextWan_2.1_ComfyUI_repackagedWan2.2-I2V-A14BRatio
4,096
8,192
16,384
32,768
65,536
131,072