Model comparison
Omni-31B-Turkish-Reasoning-Model vs Qwen3.8-27B
At Q4_K_M, Qwen3.8-27B is the smaller download — 16,810,714,336 bytes against 18,687,046,720. At long context the gap widens: Qwen3.8-27B's KV cache at 32K is 3.1× smaller, which usually matters more than the difference in weights.
From the file· summed bytes, KV per layer
Side by side
| Omni-31B-Turkish-Reasoning-Model | Qwen3.8-27B | |
|---|---|---|
| Parameters | 31.3B | 27.8B |
| Architecture | gemma4 | qwen35 |
| Layers | 60 | 64 |
| Native context | 262,144 | 262,144 |
| Mixture of experts | no | no |
| Quantizations published | 34 | 22 |
| Smallest quantization | 6.66 GiB | 8.39 GiB |
| Q4_K_M | 17.40 GiB | 15.66 GiB |
| Licence | gemma | apache-2.0 |
KV cache by context
the term that decides long-context viability
| Context | Omni-31B-Turkish-Reasoning-Model | Qwen3.8-27B | Ratio |
|---|---|---|---|
| 4,096 | 1.80 GiB | 0.25 GiB | 7.18× |
| 8,192 | 2.42 GiB | 0.50 GiB | 4.84× |
| 16,384 | 3.67 GiB | 1.00 GiB | 3.67× |
| 32,768 | 6.17 GiB | 2.00 GiB | 3.09× |
| 65,536 | 11.17 GiB | 4.00 GiB | 2.79× |
| 131,072 | 21.17 GiB | 8.00 GiB | 2.65× |