Model comparison

Qwen3.8-27B vs MN-GRAND-23.5B-Gutenberg-UNCENSORED-V2-GLM4.7-Thinking

At Q4_K_M, MN-GRAND-23.5B-Gutenberg-UNCENSORED-V2-GLM4.7-Thinking is the smaller download — 14,354,741,376 bytes against 16,810,714,336. At long context the gap widens: Qwen3.8-27B's KV cache at 32K is 5.1× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

Qwen3.8-27BMN-GRAND-23.5B-Gutenberg-UNCENSORED-V2-GLM4.7-Thinking
Parameters27.8B23.4B
Architectureqwen35llama
Layers6481
Native context262,1441,024,000
Mixture of expertsnono
Quantizations published2233
Smallest quantization8.39 GiB5.00 GiB
Q4_K_M15.66 GiB13.37 GiB
Licenceapache-2.0apache-2.0

KV cache by context

the term that decides long-context viability
ContextQwen3.8-27BMN-GRAND-23.5B-Gutenberg-UNCENSORED-V2-GLM4.7-ThinkingRatio
4,0960.25 GiB1.27 GiB5.06×
8,1920.50 GiB2.53 GiB5.06×
16,3841.00 GiB5.06 GiB5.06×
32,7682.00 GiB10.13 GiB5.06×
65,5364.00 GiB20.25 GiB5.06×
131,0728.00 GiB40.50 GiB5.06×
Qwen3.8-27B vs MN-GRAND-23.5B-Gutenberg-UNCENSORED-V2-GLM4.7-Thinking — size, memory and hardware fit — ossmodeldb