Model comparison

Gemma-The-Writer-N-Restless-Quill-10B-Uncensored vs Qwen3-8B

At Q4_K_M, Qwen3-8B is the smaller download — 5,027,783,488 bytes against 6,100,479,520. At long context the gap widens: Qwen3-8B's KV cache at 32K is 1.5× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

Gemma-The-Writer-N-Restless-Quill-10B-UncensoredQwen3-8B
Parameters10.0B8.2B
Architecturegemma2qwen3
Layers4636
Native context8,19240,960
Mixture of expertsnono
Quantizations published3151
Smallest quantization2.37 GiB2.12 GiB
Q4_K_M5.68 GiB4.68 GiB
Licenceapache-2.0apache-2.0

KV cache by context

the term that decides long-context viability
ContextGemma-The-Writer-N-Restless-Quill-10B-UncensoredQwen3-8BRatio
4,0961.44 GiB0.56 GiB2.56×
8,1922.25 GiB1.13 GiB2.00×
16,3843.68 GiB2.25 GiB1.64×
32,7686.56 GiB4.50 GiB1.46×
65,53612.31 GiB9.00 GiB1.37×
131,07223.81 GiB18.00 GiB1.32×