Model comparison

MiniMax-H3 vs WAN2.2-14B-Rapid-AllInOne

At Q4_K_M, WAN2.2-14B-Rapid-AllInOne is the smaller download — 10,849,052,416 bytes against 14,576,977,960.

From the file· summed bytes, KV per layer

Side by side

MiniMax-H3WAN2.2-14B-Rapid-AllInOne
Parameters33.1B17.3B
Architectureqwen3vlwan
Layers
Native context
Mixture of expertsnono
Quantizations published2282
Smallest quantization10.60 GiB5.59 GiB
Q4_K_M13.58 GiB10.10 GiB
Licenceother

KV cache by context

the term that decides long-context viability
ContextMiniMax-H3WAN2.2-14B-Rapid-AllInOneRatio
4,096
8,192
16,384
32,768
65,536
131,072