Model comparison

aya-expanse-8b vs Qwen3-30B-A3B-Thinking-2507

At Q4_K_M, aya-expanse-8b is the smaller download — 5,056,982,464 bytes against 18,556,685,824.

From the file· summed bytes, KV per layer

Side by side

aya-expanse-8bQwen3-30B-A3B-Thinking-2507
Parameters8.0B30.5B
Architecturecommand-rqwen3moe
Layers48
Native context262,144
Mixture of expertsnoyes, 128 experts
Quantizations published2451
Smallest quantization2.87 GiB7.05 GiB
Q4_K_M4.71 GiB17.28 GiB
Licencecc-by-nc-4.0apache-2.0

KV cache by context

the term that decides long-context viability
Contextaya-expanse-8bQwen3-30B-A3B-Thinking-2507Ratio
4,0960.38 GiB
8,1920.75 GiB
16,3841.50 GiB
32,7683.00 GiB
65,5366.00 GiB
131,07212.00 GiB