Model comparison
granite-4.0-350m vs SmolLM2-135M-Instruct
At Q4_K_M, SmolLM2-135M-Instruct is the smaller download — 105,454,144 bytes against 236,985,760. At long context the gap widens: SmolLM2-135M-Instruct's KV cache at 32K is 1.2× smaller, which usually matters more than the difference in weights.
From the file· summed bytes, KV per layer
Side by side
| granite-4.0-350m | SmolLM2-135M-Instruct | |
|---|---|---|
| Parameters | 352M | 135M |
| Architecture | granite | llama |
| Layers | 28 | 30 |
| Native context | 32,768 | 8,192 |
| Mixture of experts | no | no |
| Quantizations published | 29 | 41 |
| Smallest quantization | 0.16 GiB | 0.08 GiB |
| Q4_K_M | 0.22 GiB | 0.10 GiB |
| Licence | apache-2.0 | apache-2.0 |
KV cache by context
the term that decides long-context viability
| Context | granite-4.0-350m | SmolLM2-135M-Instruct | Ratio |
|---|---|---|---|
| 4,096 | 0.11 GiB | 0.09 GiB | 1.24× |
| 8,192 | 0.22 GiB | 0.18 GiB | 1.24× |
| 16,384 | 0.44 GiB | 0.35 GiB | 1.24× |
| 32,768 | 0.88 GiB | 0.70 GiB | 1.24× |
| 65,536 | 1.75 GiB | 1.41 GiB | 1.24× |
| 131,072 | 3.50 GiB | 2.81 GiB | 1.24× |