Model comparison
Le-Chaton-Slim-23B vs gemma-4-E4B-it
At Q4_K_M, gemma-4-E4B-it is the smaller download — 4,977,171,584 bytes against 14,183,211,648. At long context the gap widens: gemma-4-E4B-it's KV cache at 32K is 6.4× 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-E4B-it | |
|---|---|---|
| Parameters | 23.3B | 8.0B |
| Architecture | llama | gemma4 |
| Layers | 26 | 42 |
| Native context | 262,144 | 131,072 |
| Mixture of experts | yes, 10 experts | no |
| Quantizations published | 34 | 36 |
| Smallest quantization | 4.59 GiB | 3.30 GiB |
| Q4_K_M | 13.21 GiB | 4.64 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-E4B-it | Ratio |
|---|---|---|---|
| 4,096 | 0.41 GiB | 0.12 GiB | 3.30× |
| 8,192 | 0.81 GiB | 0.18 GiB | 4.57× |
| 16,384 | 1.63 GiB | 0.29 GiB | 5.66× |
| 32,768 | 3.25 GiB | 0.51 GiB | 6.42× |
| 65,536 | 6.50 GiB | 0.94 GiB | 6.89× |
| 131,072 | 13.00 GiB | 1.82 GiB | 7.15× |