Model comparison

Gemma-4-31B-Animus-V15.0 vs Qwen3-Coder-30B-A3B-Instruct

At Q4_K_M, Qwen3-Coder-30B-A3B-Instruct is the smaller download — 18,556,689,568 bytes against 18,687,064,256.

From the file· summed bytes, KV per layer

Side by side

Gemma-4-31B-Animus-V15.0Qwen3-Coder-30B-A3B-Instruct
Parameters32.7B30.5B
Architecturegemma4qwen3moe
Layers48
Native context262,144
Mixture of expertsnoyes, 128 experts
Quantizations published846
Smallest quantization11.10 GiB7.46 GiB
Q4_K_M17.40 GiB17.28 GiB
Licenceapache-2.0apache-2.0

KV cache by context

the term that decides long-context viability
ContextGemma-4-31B-Animus-V15.0Qwen3-Coder-30B-A3B-InstructRatio
4,0960.38 GiB
8,1920.75 GiB
16,3841.50 GiB
32,7683.00 GiB
65,5366.00 GiB
131,07212.00 GiB