Model comparison
Le-Chaton-Slim-23B vs gemma-4-12B-it-qat-q4_0-unquantized
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 1.3× smaller, which usually matters more than the difference in weights.
From the file· summed bytes, KV per layer
Side by side
| Le-Chaton-Slim-23B | gemma-4-12B-it-qat-q4_0-unquantized | |
|---|---|---|
| Parameters | 23.3B | 12.0B |
| Architecture | llama | gemma4 |
| Layers | 26 | 48 |
| Native context | 262,144 | 262,144 |
| Mixture of experts | yes, 10 experts | no |
| Quantizations published | 34 | 2 |
| Smallest quantization | 4.59 GiB | 6.50 GiB |
| Q4_K_M | 13.21 GiB | — |
| Licence | apache-2.0 | apache-2.0 |
KV cache by context
the term that decides long-context viability
| Context | Le-Chaton-Slim-23B | gemma-4-12B-it-qat-q4_0-unquantized | Ratio |
|---|---|---|---|
| 4,096 | 0.41 GiB | 0.72 GiB | 1.77× |
| 8,192 | 0.81 GiB | 0.97 GiB | 1.19× |
| 16,384 | 1.63 GiB | 1.47 GiB | 1.11× |
| 32,768 | 3.25 GiB | 2.47 GiB | 1.32× |
| 65,536 | 6.50 GiB | 4.47 GiB | 1.45× |
| 131,072 | 13.00 GiB | 8.47 GiB | 1.54× |