Model comparison

Qwen3-Coder-30B-A3B-Instruct vs DeepSeek-MOE-4X8B-R1-Distill-Llama-3.1-Deep-Thinker-Uncensored-24B

At Q4_K_M, DeepSeek-MOE-4X8B-R1-Distill-Llama-3.1-Deep-Thinker-Uncensored-24B is the smaller download — 15,162,181,696 bytes against 18,556,689,568.

From the file· summed bytes, KV per layer

Side by side

Qwen3-Coder-30B-A3B-InstructDeepSeek-MOE-4X8B-R1-Distill-Llama-3.1-Deep-Thinker-Uncensored-24B
Parameters30.5B24.9B
Architectureqwen3moellama
Layers48
Native context262,144
Mixture of expertsyes, 128 expertsno
Quantizations published4611
Smallest quantization7.46 GiB8.66 GiB
Q4_K_M17.28 GiB14.12 GiB
Licenceapache-2.0apache-2.0

KV cache by context

the term that decides long-context viability
ContextQwen3-Coder-30B-A3B-InstructDeepSeek-MOE-4X8B-R1-Distill-Llama-3.1-Deep-Thinker-Uncensored-24BRatio
4,0960.38 GiB
8,1920.75 GiB
16,3841.50 GiB
32,7683.00 GiB
65,5366.00 GiB
131,07212.00 GiB