Model comparison
Surogate-3.5-2B vs HyperCLOVAX-SEED-Text-Instruct-1.5B
At Q4_K_M, HyperCLOVAX-SEED-Text-Instruct-1.5B is the smaller download — 1,133,974,368 bytes against 1,598,229,728.
From the file· summed bytes, KV per layer
Side by side
| Surogate-3.5-2B | HyperCLOVAX-SEED-Text-Instruct-1.5B | |
|---|---|---|
| Parameters | 2.8B | 1.6B |
| Architecture | qwen35 | llama |
| Layers | 24 | — |
| Native context | 262,144 | — |
| Mixture of experts | no | no |
| Quantizations published | 27 | 1 |
| Smallest quantization | 1.08 GiB | 1.06 GiB |
| Q4_K_M | 1.49 GiB | 1.06 GiB |
| Licence | apache-2.0 | other |
KV cache by context
the term that decides long-context viability
| Context | Surogate-3.5-2B | HyperCLOVAX-SEED-Text-Instruct-1.5B | Ratio |
|---|---|---|---|
| 4,096 | 0.05 GiB | — | — |
| 8,192 | 0.09 GiB | — | — |
| 16,384 | 0.19 GiB | — | — |
| 32,768 | 0.38 GiB | — | — |
| 65,536 | 0.75 GiB | — | — |
| 131,072 | 1.50 GiB | — | — |