Model comparison

Wan2.2-S2V-14B vs MiniMax-H3

At Q4_K_M, Wan2.2-S2V-14B is the smaller download — 13,861,697,120 bytes against 14,576,977,960.

From the file· summed bytes, KV per layer

Side by side

Wan2.2-S2V-14BMiniMax-H3
Parameters16.3B33.1B
Architecturewanqwen3vl
Layers
Native context
Mixture of expertsnono
Quantizations published3022
Smallest quantization8.86 GiB10.60 GiB
Q4_K_M12.91 GiB13.58 GiB
Licenceapache-2.0other

KV cache by context

the term that decides long-context viability
ContextWan2.2-S2V-14BMiniMax-H3Ratio
4,096
8,192
16,384
32,768
65,536
131,072
Wan2.2-S2V-14B vs MiniMax-H3 — size, memory and hardware fit — ossmodeldb