Is LFM2-24B-A2B or Qwen3-30B-A3B-Thinking-2507 smaller?

LFM2-24B-A2B is smaller at Q4_K_M: 13.43 GiB against 17.28 GiB.

Which uses less memory at long context?

LFM2-24B-A2B — its KV cache at 32K context is 4.8x smaller, because of how its attention layers cache.

Model comparison

LFM2-24B-A2B vs Qwen3-30B-A3B-Thinking-2507

At Q4_K_M, LFM2-24B-A2B is the smaller download — 14,415,473,952 bytes against 18,556,685,824. At long context the gap widens: LFM2-24B-A2B's KV cache at 32K is 4.8× smaller, which usually matters more than the difference in weights.

From the file· summed bytes, KV per layer

Side by side

	LFM2-24B-A2B	Qwen3-30B-A3B-Thinking-2507
Parameters	23.8B	30.5B
Architecture	lfm2moe	qwen3moe
Layers	40	48
Native context	128,000	262,144
Mixture of experts	yes, 64 experts	yes, 128 experts
Quantizations published	33	51
Smallest quantization	5.35 GiB	7.05 GiB
Q4_K_M	13.43 GiB	17.28 GiB
Licence	other	apache-2.0

KV cache by context

the term that decides long-context viability

Context	LFM2-24B-A2B	Qwen3-30B-A3B-Thinking-2507	Ratio
4,096	0.08 GiB	0.38 GiB	4.80×
8,192	0.16 GiB	0.75 GiB	4.80×
16,384	0.31 GiB	1.50 GiB	4.80×
32,768	0.63 GiB	3.00 GiB	4.80×
65,536	1.25 GiB	6.00 GiB	4.80×
131,072	2.50 GiB	12.00 GiB	4.80×