Model comparison

MiniMax-H3 vs Wan14BT2VFusioniX

At Q4_K_M, Wan14BT2VFusioniX is the smaller download — 9,650,100,736 bytes against 14,576,977,960.

From the file· summed bytes, KV per layer

Side by side

MiniMax-H3Wan14BT2VFusioniX
Parameters33.1B14.3B
Architectureqwen3vlwan
Layers
Native context
Mixture of expertsnono
Quantizations published2229
Smallest quantization10.60 GiB4.94 GiB
Q4_K_M13.58 GiB8.99 GiB
Licenceotherapache-2.0

KV cache by context

the term that decides long-context viability
ContextMiniMax-H3Wan14BT2VFusioniXRatio
4,096
8,192
16,384
32,768
65,536
131,072