Model comparison

Qwen3-Coder-30B-A3B-Instruct vs EXAONE-4.0-32B

At Q4_K_M, Qwen3-Coder-30B-A3B-Instruct is the smaller download — 18,556,689,568 bytes against 19,343,826,720.

From the file· summed bytes, KV per layer

Side by side

Qwen3-Coder-30B-A3B-InstructEXAONE-4.0-32B
Parameters30.5B32.0B
Architectureqwen3moeexaone4
Layers4864
Native context262,144131,072
Mixture of expertsyes, 128 expertsno
Quantizations published4631
Smallest quantization7.46 GiB8.96 GiB
Q4_K_M17.28 GiB18.02 GiB
Licenceapache-2.0other

KV cache by context

the term that decides long-context viability
ContextQwen3-Coder-30B-A3B-InstructEXAONE-4.0-32BRatio
4,0960.38 GiB1.00 GiB2.67×
8,1920.75 GiB1.34 GiB1.79×
16,3841.50 GiB1.84 GiB1.23×
32,7683.00 GiB2.84 GiB1.05×
65,5366.00 GiB4.84 GiB1.24×
131,07212.00 GiB8.84 GiB1.36×