Model comparison

internlm3-8b-instruct vs Qwen3-8B

At Q4_K_M, Qwen3-8B is the smaller download — 5,027,783,488 bytes against 5,358,623,936. At long context the gap widens: internlm3-8b-instruct's KV cache at 32K is 3.0× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

internlm3-8b-instructQwen3-8B
Parameters8.8B8.2B
Architecturellamaqwen3
Layers4836
Native context32,76840,960
Mixture of expertsnono
Quantizations published3551
Smallest quantization2.98 GiB2.12 GiB
Q4_K_M4.99 GiB4.68 GiB
Licenceapache-2.0apache-2.0

KV cache by context

the term that decides long-context viability
Contextinternlm3-8b-instructQwen3-8BRatio
4,0960.19 GiB0.56 GiB3.00×
8,1920.38 GiB1.13 GiB3.00×
16,3840.75 GiB2.25 GiB3.00×
32,7681.50 GiB4.50 GiB3.00×
65,5363.00 GiB9.00 GiB3.00×
131,0726.00 GiB18.00 GiB3.00×