Model comparison
gemma-4-12B-it vs QwenPaw-Flash-9B-heretic
At Q4_K_M, QwenPaw-Flash-9B-heretic is the smaller download — 5,629,109,184 bytes against 7,121,861,440. At long context the gap widens: QwenPaw-Flash-9B-heretic's KV cache at 32K is 2.5× smaller, which usually matters more than the difference in weights.
From the file· summed bytes, KV per layer
Side by side
| gemma-4-12B-it | QwenPaw-Flash-9B-heretic | |
|---|---|---|
| Parameters | 12.0B | 9.4B |
| Architecture | gemma4 | clip |
| Layers | 48 | 32 |
| Native context | 262,144 | 262,144 |
| Mixture of experts | no | no |
| Quantizations published | 59 | 6 |
| Smallest quantization | 3.92 GiB | 2.02 GiB |
| Q4_K_M | 6.63 GiB | 5.24 GiB |
| Licence | apache-2.0 | apache-2.0 |
KV cache by context
the term that decides long-context viability
| Context | gemma-4-12B-it | QwenPaw-Flash-9B-heretic | Ratio |
|---|---|---|---|
| 4,096 | 0.72 GiB | 0.13 GiB | 5.75× |
| 8,192 | 0.97 GiB | 0.25 GiB | 3.87× |
| 16,384 | 1.47 GiB | 0.50 GiB | 2.94× |
| 32,768 | 2.47 GiB | 1.00 GiB | 2.47× |
| 65,536 | 4.47 GiB | 2.00 GiB | 2.23× |
| 131,072 | 8.47 GiB | 4.00 GiB | 2.12× |