Model comparison

gemma-4-12B-it vs Qwen3-VL-8B-Instruct-abliterated-v2

At Q4_K_M, Qwen3-VL-8B-Instruct-abliterated-v2 is the smaller download — 5,027,785,824 bytes against 7,121,861,440.

From the file· summed bytes, KV per layer

Side by side

gemma-4-12B-itQwen3-VL-8B-Instruct-abliterated-v2
Parameters12.0B8.8B
Architecturegemma4qwen3vl
Layers48
Native context262,144
Mixture of expertsnono
Quantizations published5936
Smallest quantization3.92 GiB3.06 GiB
Q4_K_M6.63 GiB4.68 GiB
Licenceapache-2.0apache-2.0

KV cache by context

the term that decides long-context viability
Contextgemma-4-12B-itQwen3-VL-8B-Instruct-abliterated-v2Ratio
4,0960.72 GiB
8,1920.97 GiB
16,3841.47 GiB
32,7682.47 GiB
65,5364.47 GiB
131,0728.47 GiB
gemma-4-12B-it vs Qwen3-VL-8B-Instruct-abliterated-v2 — size, memory and hardware fit — ossmodeldb