Model comparison

DeepSeek-Coder-V2-Instruct-0724 vs Hy3

At Q4_K_M, DeepSeek-Coder-V2-Instruct-0724 is the smaller download — 142,453,966,944 bytes against 180,999,059,328. At long context the gap widens: DeepSeek-Coder-V2-Instruct-0724's KV cache at 32K is 4.7× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

DeepSeek-Coder-V2-Instruct-0724Hy3
Parameters236B299B
Architecturedeepseek2hy_v3
Layers6080
Native context163,840262,144
Mixture of expertsyes, 160 expertsyes, 192 experts
Quantizations published4437
Smallest quantization44.14 GiB59.65 GiB
Q4_K_M132.67 GiB168.57 GiB
Licenceotherapache-2.0

KV cache by context

the term that decides long-context viability
ContextDeepSeek-Coder-V2-Instruct-0724Hy3Ratio
4,0960.26 GiB1.25 GiB4.74×
8,1920.53 GiB2.50 GiB4.74×
16,3841.05 GiB5.00 GiB4.74×
32,7682.11 GiB10.00 GiB4.74×
65,5364.22 GiB20.00 GiB4.74×
131,0728.44 GiB40.00 GiB4.74×
DeepSeek-Coder-V2-Instruct-0724 vs Hy3 — size, memory and hardware fit — ossmodeldb