Model comparison

HyperCLOVAX-SEED-Text-Instruct-1.5B vs Qwen2.5-7B-Instruct

At Q4_K_M, HyperCLOVAX-SEED-Text-Instruct-1.5B is the smaller download — 1,133,974,368 bytes against 4,683,073,632.

From the file· summed bytes, KV per layer

Side by side

HyperCLOVAX-SEED-Text-Instruct-1.5BQwen2.5-7B-Instruct
Parameters1.6B7.6B
Architecturellamaqwen2
Layers28
Native context32,768
Mixture of expertsnono
Quantizations published132
Smallest quantization1.06 GiB2.59 GiB
Q4_K_M1.06 GiB4.36 GiB
Licenceotherapache-2.0

KV cache by context

the term that decides long-context viability
ContextHyperCLOVAX-SEED-Text-Instruct-1.5BQwen2.5-7B-InstructRatio
4,0960.22 GiB
8,1920.44 GiB
16,3840.88 GiB
32,7681.75 GiB
65,5363.50 GiB
131,0727.00 GiB