Model comparison
medgemma-27b-it vs Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-MTP
At Q4_K_M, medgemma-27b-it is the smaller download — 16,546,689,536 bytes against 36,545,830,912. At long context the gap widens: Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-MTP's KV cache at 32K is 1.6× smaller, which usually matters more than the difference in weights.
From the file· summed bytes, KV per layer
Side by side
| medgemma-27b-it | Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-MTP | |
|---|---|---|
| Parameters | 28.8B | 27.8B |
| Architecture | gemma3 | qwen35 |
| Layers | 62 | 64 |
| Native context | 131,072 | 262,144 |
| Mixture of experts | no | no |
| Quantizations published | 68 | 34 |
| Smallest quantization | 5.83 GiB | 10.12 GiB |
| Q4_K_M | 15.41 GiB | 34.04 GiB |
| Licence | other | apache-2.0 |
KV cache by context
the term that decides long-context viability
| Context | medgemma-27b-it | Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-MTP | Ratio |
|---|---|---|---|
| 4,096 | 0.92 GiB | 0.25 GiB | 3.68× |
| 8,192 | 1.23 GiB | 0.50 GiB | 2.47× |
| 16,384 | 1.86 GiB | 1.00 GiB | 1.86× |
| 32,768 | 3.11 GiB | 2.00 GiB | 1.55× |
| 65,536 | 5.61 GiB | 4.00 GiB | 1.40× |
| 131,072 | 10.61 GiB | 8.00 GiB | 1.33× |