Model comparison

swarm-sovereign-12b vs Qwen3-8B

These two publish different quantization sets; the table below has the exact sizes. At long context the gap widens: swarm-sovereign-12b's KV cache at 32K is 1.8× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

swarm-sovereign-12bQwen3-8B
Parameters12.0B8.2B
Architecturegemma4qwen3
Layers4836
Native context131,07240,960
Mixture of expertsnono
Quantizations published2451
Smallest quantization2.78 GiB2.12 GiB
Q4_K_M4.68 GiB
Licenceapache-2.0apache-2.0

KV cache by context

the term that decides long-context viability
Contextswarm-sovereign-12bQwen3-8BRatio
4,0960.72 GiB0.56 GiB1.28×
8,1920.97 GiB1.13 GiB1.16×
16,3841.47 GiB2.25 GiB1.53×
32,7682.47 GiB4.50 GiB1.82×
65,5364.47 GiB9.00 GiB2.01×
131,0728.47 GiB18.00 GiB2.13×