Model comparison
gemma-4-12B-it vs MR_MARTIN_31b_V1.0_Gemma-4-base-ftp
These two publish different quantization sets; the table below has the exact sizes.
From the file· summed bytes, KV per layer
Side by side
| gemma-4-12B-it | MR_MARTIN_31b_V1.0_Gemma-4-base-ftp | |
|---|---|---|
| Parameters | 12.0B | 30.7B |
| Architecture | gemma4 | gemma4 |
| Layers | 48 | — |
| Native context | 262,144 | — |
| Mixture of experts | no | no |
| Quantizations published | 59 | 23 |
| Smallest quantization | 3.92 GiB | 6.66 GiB |
| Q4_K_M | 6.63 GiB | — |
| Licence | apache-2.0 | other |
KV cache by context
the term that decides long-context viability
| Context | gemma-4-12B-it | MR_MARTIN_31b_V1.0_Gemma-4-base-ftp | Ratio |
|---|---|---|---|
| 4,096 | 0.72 GiB | — | — |
| 8,192 | 0.97 GiB | — | — |
| 16,384 | 1.47 GiB | — | — |
| 32,768 | 2.47 GiB | — | — |
| 65,536 | 4.47 GiB | — | — |
| 131,072 | 8.47 GiB | — | — |