Model comparison
gemma-3-12b-it-qat-int4-unquantized vs Qwen3.6-35B-A3B
At Q4_K_M, gemma-3-12b-it-qat-int4-unquantized is the smaller download — 7,300,778,848 bytes against 21,166,757,728.
From the file· summed bytes, KV per layer
Side by side
| gemma-3-12b-it-qat-int4-unquantized | Qwen3.6-35B-A3B | |
|---|---|---|
| Parameters | 12.2B | 36.0B |
| Architecture | gemma3 | qwen35moe |
| Layers | — | 40 |
| Native context | — | 262,144 |
| Mixture of experts | no | yes, 256 experts |
| Quantizations published | 15 | 61 |
| Smallest quantization | 4.11 GiB | 8.77 GiB |
| Q4_K_M | 6.80 GiB | 19.71 GiB |
| Licence | gemma | apache-2.0 |
KV cache by context
the term that decides long-context viability
| Context | gemma-3-12b-it-qat-int4-unquantized | Qwen3.6-35B-A3B | Ratio |
|---|---|---|---|
| 4,096 | — | 0.08 GiB | — |
| 8,192 | — | 0.16 GiB | — |
| 16,384 | — | 0.31 GiB | — |
| 32,768 | — | 0.63 GiB | — |
| 65,536 | — | 1.25 GiB | — |
| 131,072 | — | 2.50 GiB | — |