Model comparison

Mistral-Small-3.1-24B-Instruct-2503 vs Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-MTP

At Q4_K_M, Mistral-Small-3.1-24B-Instruct-2503 is the smaller download — 14,333,910,176 bytes against 36,545,830,912. At long context the gap widens: Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-MTP's KV cache at 32K is 2.5× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

Mistral-Small-3.1-24B-Instruct-2503Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-MTP
Parameters24.0B27.8B
Architecturellamaqwen35
Layers4064
Native context131,072262,144
Mixture of expertsnono
Quantizations published5034
Smallest quantization5.18 GiB10.12 GiB
Q4_K_M13.35 GiB34.04 GiB
Licenceapache-2.0apache-2.0

KV cache by context

the term that decides long-context viability
ContextMistral-Small-3.1-24B-Instruct-2503Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-MTPRatio
4,0960.63 GiB0.25 GiB2.50×
8,1921.25 GiB0.50 GiB2.50×
16,3842.50 GiB1.00 GiB2.50×
32,7685.00 GiB2.00 GiB2.50×
65,53610.00 GiB4.00 GiB2.50×
131,07220.00 GiB8.00 GiB2.50×