Model comparison
gemma-4-12B-it-qat-q4_0-unquantized vs Olmo-3.1-32B-Instruct
These two publish different quantization sets; the table below has the exact sizes. At long context the gap widens: gemma-4-12B-it-qat-q4_0-unquantized's KV cache at 32K is 1.2× 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-qat-q4_0-unquantized | Olmo-3.1-32B-Instruct | |
|---|---|---|
| Parameters | 12.0B | 32.2B |
| Architecture | gemma4 | olmo2 |
| Layers | 48 | 64 |
| Native context | 262,144 | 65,536 |
| Mixture of experts | no | no |
| Quantizations published | 2 | 49 |
| Smallest quantization | 6.50 GiB | 6.75 GiB |
| Q4_K_M | — | 18.14 GiB |
| Licence | apache-2.0 | apache-2.0 |
KV cache by context
the term that decides long-context viability
| Context | gemma-4-12B-it-qat-q4_0-unquantized | Olmo-3.1-32B-Instruct | Ratio |
|---|---|---|---|
| 4,096 | 0.72 GiB | 1.00 GiB | 1.39× |
| 8,192 | 0.97 GiB | 1.34 GiB | 1.39× |
| 16,384 | 1.47 GiB | 1.84 GiB | 1.26× |
| 32,768 | 2.47 GiB | 2.84 GiB | 1.15× |
| 65,536 | 4.47 GiB | 4.84 GiB | 1.08× |
| 131,072 | 8.47 GiB | 8.84 GiB | 1.04× |