Model comparison

Qwen3-Coder-30B-A3B-Instruct vs Llama-3.2-8X3B-GATED-MOE-Reasoning-Dark-Champion-Instruct-uncensored-abliterated-18.4B

At Q4_K_M, Llama-3.2-8X3B-GATED-MOE-Reasoning-Dark-Champion-Instruct-uncensored-abliterated-18.4B is the smaller download — 11,408,360,768 bytes against 18,556,689,568.

From the file· summed bytes, KV per layer

Side by side

Qwen3-Coder-30B-A3B-InstructLlama-3.2-8X3B-GATED-MOE-Reasoning-Dark-Champion-Instruct-uncensored-abliterated-18.4B
Parameters30.5B18.4B
Architectureqwen3moellama
Layers48
Native context262,144
Mixture of expertsyes, 128 expertsno
Quantizations published4611
Smallest quantization7.46 GiB6.65 GiB
Q4_K_M17.28 GiB10.62 GiB
Licenceapache-2.0apache-2.0

KV cache by context

the term that decides long-context viability
ContextQwen3-Coder-30B-A3B-InstructLlama-3.2-8X3B-GATED-MOE-Reasoning-Dark-Champion-Instruct-uncensored-abliterated-18.4BRatio
4,0960.38 GiB
8,1920.75 GiB
16,3841.50 GiB
32,7683.00 GiB
65,5366.00 GiB
131,07212.00 GiB