Model comparison

JoyAI-Echo vs MiniMax-H3

At Q4_K_M, MiniMax-H3 is the smaller download — 14,576,977,960 bytes against 15,694,613,568.

From the file· summed bytes, KV per layer

Side by side

JoyAI-EchoMiniMax-H3
Parameters12.2B33.1B
Architecturejoyai-gemmaqwen3vl
Layers
Native context
Mixture of expertsnono
Quantizations published522
Smallest quantization8.54 GiB10.60 GiB
Q4_K_M14.62 GiB13.58 GiB
Licenceother

KV cache by context

the term that decides long-context viability
ContextJoyAI-EchoMiniMax-H3Ratio
4,096
8,192
16,384
32,768
65,536
131,072