Model comparison

gemma-4-12B-it vs Qwen2.5-VL-7B-NSFW-Caption-V3

At Q4_K_M, Qwen2.5-VL-7B-NSFW-Caption-V3 is the smaller download — 4,683,072,384 bytes against 7,121,861,440.

From the file· summed bytes, KV per layer

Side by side

gemma-4-12B-itQwen2.5-VL-7B-NSFW-Caption-V3
Parameters12.0B7.6B
Architecturegemma4qwen2vl
Layers48
Native context262,144
Mixture of expertsnono
Quantizations published5958
Smallest quantization3.92 GiB2.59 GiB
Q4_K_M6.63 GiB4.36 GiB
Licenceapache-2.0apache-2.0

KV cache by context

the term that decides long-context viability
Contextgemma-4-12B-itQwen2.5-VL-7B-NSFW-Caption-V3Ratio
4,0960.72 GiB
8,1920.97 GiB
16,3841.47 GiB
32,7682.47 GiB
65,5364.47 GiB
131,0728.47 GiB