Is ultravox-v0_5-llama-3_2-1b or Qwen3-1.7B smaller?

ultravox-v0_5-llama-3_2-1b is smaller at Q4_K_M: 0.75 GiB against 1.03 GiB.

Which uses less memory at long context?

Qwen3-1.7B — its KV cache at 32K context is 2.3x smaller, because of how its attention layers cache.

Model comparison

ultravox-v0_5-llama-3_2-1b vs Qwen3-1.7B

At Q4_K_M, ultravox-v0_5-llama-3_2-1b is the smaller download — 807,694,464 bytes against 1,107,409,472. At long context the gap widens: Qwen3-1.7B's KV cache at 32K is 2.3× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	ultravox-v0_5-llama-3_2-1b	Qwen3-1.7B
Parameters	683M	2.0B
Architecture	llama	qwen3
Layers	16	28
Native context	—	40,960
Mixture of experts	no	no
Quantizations published	2	48
Smallest quantization	0.75 GiB	0.50 GiB
Q4_K_M	0.75 GiB	1.03 GiB
Licence	mit	—

KV cache by context

the term that decides long-context viability

Context	ultravox-v0_5-llama-3_2-1b	Qwen3-1.7B	Ratio
4,096	1.00 GiB	0.44 GiB	2.29×
8,192	2.00 GiB	0.88 GiB	2.29×
16,384	4.00 GiB	1.75 GiB	2.29×
32,768	8.00 GiB	3.50 GiB	2.29×
65,536	16.00 GiB	7.00 GiB	2.29×
131,072	32.00 GiB	14.00 GiB	2.29×