Model comparison

Qwen3.6-27B vs Mistral-Small-3.2-24B-Instruct-2507

At Q4_K_M, Mistral-Small-3.2-24B-Instruct-2507 is the smaller download — 14,333,910,048 bytes against 16,547,397,888.

From the file· summed bytes, KV per layer

Side by side

Qwen3.6-27BMistral-Small-3.2-24B-Instruct-2507
Parameters27.8B23.6B
Architectureqwen35llama
Layers64
Native context262,144
Mixture of expertsnono
Quantizations published4018
Smallest quantization8.74 GiB8.28 GiB
Q4_K_M15.41 GiB13.35 GiB
Licenceapache-2.0

KV cache by context

the term that decides long-context viability
ContextQwen3.6-27BMistral-Small-3.2-24B-Instruct-2507Ratio
4,0960.25 GiB
8,1920.50 GiB
16,3841.00 GiB
32,7682.00 GiB
65,5364.00 GiB
131,0728.00 GiB