Is SmolLM2-1.7B-Instruct-Uncensored or HyperCLOVAX-SEED-Text-Instruct-1.5B smaller?

SmolLM2-1.7B-Instruct-Uncensored is smaller at Q4_K_M: 1.04 GiB against 1.06 GiB.

Which uses less memory at long context?

We cannot resolve the KV layout for one of these models, so we do not publish a comparison.

Model comparison

At Q4_K_M, SmolLM2-1.7B-Instruct-Uncensored is the smaller download — 1,112,233,344 bytes against 1,133,974,368.

From the file· summed bytes, KV per layer

	SmolLM2-1.7B-Instruct-Uncensored	HyperCLOVAX-SEED-Text-Instruct-1.5B
Parameters	1.8B	1.6B
Architecture	llama	llama
Layers	24	—
Native context	8,192	—
Mixture of experts	no	no
Quantizations published	12	1
Smallest quantization	0.66 GiB	1.06 GiB
Q4_K_M	1.04 GiB	1.06 GiB
Licence	apache-2.0	other

the term that decides long-context viability

Context	SmolLM2-1.7B-Instruct-Uncensored	HyperCLOVAX-SEED-Text-Instruct-1.5B	Ratio
4,096	0.75 GiB	—	—
8,192	1.50 GiB	—	—
16,384	3.00 GiB	—	—
32,768	6.00 GiB	—	—
65,536	12.00 GiB	—	—
131,072	24.00 GiB	—	—