Model comparison

Qwen3.6-27B vs Qwen3.6-12B-IQ-Ultra-Heretic-Uncensored-Thinking-V2-Hightop

At Q4_K_M, Qwen3.6-12B-IQ-Ultra-Heretic-Uncensored-Thinking-V2-Hightop is the smaller download — 7,323,108,608 bytes against 16,547,397,888. At long context the gap widens: Qwen3.6-12B-IQ-Ultra-Heretic-Uncensored-Thinking-V2-Hightop's KV cache at 32K is 2.7× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

Qwen3.6-27BQwen3.6-12B-IQ-Ultra-Heretic-Uncensored-Thinking-V2-Hightop
Parameters27.8B12.1B
Architectureqwen35qwen35
Layers6424
Native context262,144262,144
Mixture of expertsnono
Quantizations published4033
Smallest quantization8.74 GiB4.60 GiB
Q4_K_M15.41 GiB6.82 GiB
Licenceapache-2.0apache-2.0

KV cache by context

the term that decides long-context viability
ContextQwen3.6-27BQwen3.6-12B-IQ-Ultra-Heretic-Uncensored-Thinking-V2-HightopRatio
4,0960.25 GiB0.09 GiB2.67×
8,1920.50 GiB0.19 GiB2.67×
16,3841.00 GiB0.38 GiB2.67×
32,7682.00 GiB0.75 GiB2.67×
65,5364.00 GiB1.50 GiB2.67×
131,0728.00 GiB3.00 GiB2.67×