Is Qwen3.8-27B or Trinity-Mini smaller?

Trinity-Mini is smaller at Q4_K_M: 14.74 GiB against 15.66 GiB.

Which uses less memory at long context?

Trinity-Mini — its KV cache at 32K context is 3.2x smaller, because of how its attention layers cache.

Model comparison

Qwen3.8-27B vs Trinity-Mini

At Q4_K_M, Trinity-Mini is the smaller download — 15,823,053,440 bytes against 16,810,714,336. At long context the gap widens: Trinity-Mini's KV cache at 32K is 3.2× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	Qwen3.8-27B	Trinity-Mini
Parameters	27.8B	26.1B
Architecture	qwen35	afmoe
Layers	64	32
Native context	262,144	131,072
Mixture of experts	no	yes, 128 experts
Quantizations published	22	57
Smallest quantization	8.39 GiB	6.11 GiB
Q4_K_M	15.66 GiB	14.74 GiB
Licence	apache-2.0	apache-2.0

KV cache by context

the term that decides long-context viability

Context	Qwen3.8-27B	Trinity-Mini	Ratio
4,096	0.25 GiB	0.18 GiB	1.39×
8,192	0.50 GiB	0.24 GiB	2.07×
16,384	1.00 GiB	0.37 GiB	2.72×
32,768	2.00 GiB	0.62 GiB	3.24×
65,536	4.00 GiB	1.12 GiB	3.58×
131,072	8.00 GiB	2.12 GiB	3.78×