Model comparison
Bielik-11B-v2.3-Instruct vs Qwen3.8-27B
At Q4_K_M, Bielik-11B-v2.3-Instruct is the smaller download — 6,724,050,496 bytes against 16,810,714,336. At long context the gap widens: Qwen3.8-27B's KV cache at 32K is 3.1× smaller, which usually matters more than the difference in weights.
From the file· summed bytes, KV per layer
Side by side
| Bielik-11B-v2.3-Instruct | Qwen3.8-27B | |
|---|---|---|
| Parameters | 11.2B | 27.8B |
| Architecture | llama | qwen35 |
| Layers | 50 | 64 |
| Native context | 32,768 | 262,144 |
| Mixture of experts | no | no |
| Quantizations published | 4 | 22 |
| Smallest quantization | 6.26 GiB | 8.39 GiB |
| Q4_K_M | 6.26 GiB | 15.66 GiB |
| Licence | apache-2.0 | apache-2.0 |
KV cache by context
the term that decides long-context viability
| Context | Bielik-11B-v2.3-Instruct | Qwen3.8-27B | Ratio |
|---|---|---|---|
| 4,096 | 0.78 GiB | 0.25 GiB | 3.12× |
| 8,192 | 1.56 GiB | 0.50 GiB | 3.12× |
| 16,384 | 3.13 GiB | 1.00 GiB | 3.12× |
| 32,768 | 6.25 GiB | 2.00 GiB | 3.12× |
| 65,536 | 12.50 GiB | 4.00 GiB | 3.12× |
| 131,072 | 25.00 GiB | 8.00 GiB | 3.12× |