Model comparison

starcoder2-7b vs Qwen3-8B

At Q4_K_M, starcoder2-7b is the smaller download — 4,402,887,488 bytes against 5,027,783,488.

From the file· summed bytes, KV per layer

Side by side

starcoder2-7bQwen3-8B
Parameters7.2B8.2B
Architecturestarcoder2qwen3
Layers3236
Native context16,38440,960
Mixture of expertsnono
Quantizations published1251
Smallest quantization2.54 GiB2.12 GiB
Q4_K_M4.10 GiB4.68 GiB
Licencebigcode-openrail-mapache-2.0

KV cache by context

the term that decides long-context viability
Contextstarcoder2-7bQwen3-8BRatio
4,0960.56 GiB
8,1921.13 GiB
16,3842.25 GiB
32,7684.50 GiB
65,5369.00 GiB
131,07218.00 GiB