Is orpheus-3b-0.1-ft or neutts-air smaller?

neutts-air is smaller at Q4_K_M: 0.43 GiB against 1.95 GiB.

Which uses less memory at long context?

neutts-air — its KV cache at 32K context is 9.3x smaller, because of how its attention layers cache.

Model comparison

orpheus-3b-0.1-ft vs neutts-air

At Q4_K_M, neutts-air is the smaller download — 462,193,248 bytes against 2,092,567,712. At long context the gap widens: neutts-air's KV cache at 32K is 9.3× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	orpheus-3b-0.1-ft	neutts-air
Parameters	3.8B	748M
Architecture	llama	qwen2
Layers	28	24
Native context	131,072	32,768
Mixture of experts	no	no
Quantizations published	22	6
Smallest quantization	0.91 GiB	0.43 GiB
Q4_K_M	1.95 GiB	0.43 GiB
Licence	apache-2.0	apache-2.0

KV cache by context

the term that decides long-context viability

Context	orpheus-3b-0.1-ft	neutts-air	Ratio
4,096	0.44 GiB	0.05 GiB	9.33×
8,192	0.88 GiB	0.09 GiB	9.33×
16,384	1.75 GiB	0.19 GiB	9.33×
32,768	3.50 GiB	0.38 GiB	9.33×
65,536	7.00 GiB	0.75 GiB	9.33×
131,072	14.00 GiB	1.50 GiB	9.33×