Is Qwen3-Coder-30B-A3B-Instruct or Delphi-25B-SimpleRL-Math smaller?

Delphi-25B-SimpleRL-Math is smaller at Q4_K_M: 14.12 GiB against 17.28 GiB.

Which uses less memory at long context?

Qwen3-Coder-30B-A3B-Instruct — its KV cache at 32K context is 11.2x smaller, because of how its attention layers cache.

Model comparison

Qwen3-Coder-30B-A3B-Instruct vs Delphi-25B-SimpleRL-Math

At Q4_K_M, Delphi-25B-SimpleRL-Math is the smaller download — 15,160,698,624 bytes against 18,556,689,568. At long context the gap widens: Qwen3-Coder-30B-A3B-Instruct's KV cache at 32K is 11.2× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	Qwen3-Coder-30B-A3B-Instruct	Delphi-25B-SimpleRL-Math
Parameters	30.5B	25.0B
Architecture	qwen3moe	qwen3
Layers	48	51
Native context	262,144	16,384
Mixture of experts	yes, 128 experts	no
Quantizations published	46	34
Smallest quantization	7.46 GiB	5.31 GiB
Q4_K_M	17.28 GiB	14.12 GiB
Licence	apache-2.0	—

KV cache by context

the term that decides long-context viability

Context	Qwen3-Coder-30B-A3B-Instruct	Delphi-25B-SimpleRL-Math	Ratio
4,096	0.38 GiB	4.18 GiB	11.16×
8,192	0.75 GiB	8.37 GiB	11.16×
16,384	1.50 GiB	16.73 GiB	11.16×
32,768	3.00 GiB	33.47 GiB	11.16×
65,536	6.00 GiB	66.94 GiB	11.16×
131,072	12.00 GiB	133.88 GiB	11.16×