Model comparison

OLMoE-1B-7B-0924-Instruct vs Qwen3-8B

At Q4_K_M, OLMoE-1B-7B-0924-Instruct is the smaller download — 4,213,512,672 bytes against 5,027,783,488.

From the file· summed bytes, KV per layer

Side by side

OLMoE-1B-7B-0924-InstructQwen3-8B
Parameters6.9B8.2B
Architectureolmoeqwen3
Layers1636
Native context4,09640,960
Mixture of expertsyes, 64 expertsno
Quantizations published5451
Smallest quantization1.39 GiB2.12 GiB
Q4_K_M3.92 GiB4.68 GiB
Licenceapache-2.0apache-2.0

KV cache by context

the term that decides long-context viability
ContextOLMoE-1B-7B-0924-InstructQwen3-8BRatio
4,0960.50 GiB0.56 GiB1.13×
8,1921.00 GiB1.13 GiB1.13×
16,3842.00 GiB2.25 GiB1.13×
32,7684.00 GiB4.50 GiB1.13×
65,5368.00 GiB9.00 GiB1.13×
131,07216.00 GiB18.00 GiB1.13×