Model comparison

Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-MTP vs Qwen2.5-VL-72B-Instruct

At Q4_K_M, Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-MTP is the smaller download — 36,545,830,912 bytes against 47,415,713,248. At long context the gap widens: Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-MTP's KV cache at 32K is 5.0× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-MTPQwen2.5-VL-72B-Instruct
Parameters27.8B73.4B
Architectureqwen35qwen2vl
Layers6480
Native context262,144128,000
Mixture of expertsnono
Quantizations published3445
Smallest quantization10.12 GiB22.11 GiB
Q4_K_M34.04 GiB44.16 GiB
Licenceapache-2.0other

KV cache by context

the term that decides long-context viability
ContextQwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-MTPQwen2.5-VL-72B-InstructRatio
4,0960.25 GiB1.25 GiB5.00×
8,1920.50 GiB2.50 GiB5.00×
16,3841.00 GiB5.00 GiB5.00×
32,7682.00 GiB10.00 GiB5.00×
65,5364.00 GiB20.00 GiB5.00×
131,0728.00 GiB40.00 GiB5.00×