Model comparison

InternVL3_5-30B-A3B vs gemma-4-26B-A4B-it

At Q4_K_M, gemma-4-26B-A4B-it is the smaller download — 16,796,017,248 bytes against 18,632,180,352.

From the file· summed bytes, KV per layer

Side by side

InternVL3_5-30B-A3Bgemma-4-26B-A4B-it
Parameters30.8B26.5B
Architectureqwen3moegemma4
Layers30
Native context262,144
Mixture of expertsnoyes, 128 experts
Quantizations published3046
Smallest quantization7.05 GiB8.99 GiB
Q4_K_M17.35 GiB15.64 GiB
Licenceapache-2.0

KV cache by context

the term that decides long-context viability
ContextInternVL3_5-30B-A3Bgemma-4-26B-A4B-itRatio
4,0960.45 GiB
8,1920.61 GiB
16,3840.92 GiB
32,7681.54 GiB
65,5362.79 GiB
131,0725.29 GiB