Model comparison
Llama-3.2-4X3B-MOE-Hell-California-10B vs gemma-4-12B-it-qat-q4_0-unquantized
These two publish different quantization sets; the table below has the exact sizes.
From the file· summed bytes, KV per layer
Side by side
| Llama-3.2-4X3B-MOE-Hell-California-10B | gemma-4-12B-it-qat-q4_0-unquantized | |
|---|---|---|
| Parameters | 9.9B | 12.0B |
| Architecture | llama | gemma4 |
| Layers | — | 48 |
| Native context | — | 262,144 |
| Mixture of experts | no | no |
| Quantizations published | 11 | 2 |
| Smallest quantization | 3.53 GiB | 6.50 GiB |
| Q4_K_M | 5.66 GiB | — |
| Licence | apache-2.0 | apache-2.0 |
KV cache by context
the term that decides long-context viability
| Context | Llama-3.2-4X3B-MOE-Hell-California-10B | gemma-4-12B-it-qat-q4_0-unquantized | 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 | — |