Model comparison

Qwen3-1.7B-Distilled-30B-A3B-SFT vs Llama-3.1-8B-Instruct

At Q4_K_M, Qwen3-1.7B-Distilled-30B-A3B-SFT is the smaller download — 1,282,439,072 bytes against 4,920,739,168.

From the file· summed bytes, KV per layer

Side by side

Qwen3-1.7B-Distilled-30B-A3B-SFTLlama-3.1-8B-Instruct
Parameters2.0B8.0B
Architectureqwen3llama
Layers2832
Native context40,960131,072
Mixture of expertsnono
Quantizations published445
Smallest quantization1.19 GiB2.02 GiB
Q4_K_M1.19 GiB4.58 GiB
Licenceapache-2.0llama3.1

KV cache by context

the term that decides long-context viability
ContextQwen3-1.7B-Distilled-30B-A3B-SFTLlama-3.1-8B-InstructRatio
4,0960.44 GiB0.50 GiB1.14×
8,1920.88 GiB1.00 GiB1.14×
16,3841.75 GiB2.00 GiB1.14×
32,7683.50 GiB4.00 GiB1.14×
65,5367.00 GiB8.00 GiB1.14×
131,07214.00 GiB16.00 GiB1.14×