Model comparison
Qwen3.6-27B vs Qwen3.6-27B-NVFP4
These two publish different quantization sets; the table below has the exact sizes.
From the file· summed bytes, KV per layer
Side by side
| Qwen3.6-27B | Qwen3.6-27B-NVFP4 | |
|---|---|---|
| Parameters | 27.8B | 18.2B |
| Architecture | qwen35 | qwen35 |
| Layers | 64 | 64 |
| Native context | 262,144 | 262,144 |
| Mixture of experts | no | no |
| Quantizations published | 40 | 2 |
| Smallest quantization | 8.74 GiB | 18.51 GiB |
| Q4_K_M | 15.41 GiB | — |
| Licence | apache-2.0 | apache-2.0 |
KV cache by context
the term that decides long-context viability
| Context | Qwen3.6-27B | Qwen3.6-27B-NVFP4 | Ratio |
|---|---|---|---|
| 4,096 | 0.25 GiB | 0.25 GiB | 1.00× |
| 8,192 | 0.50 GiB | 0.50 GiB | 1.00× |
| 16,384 | 1.00 GiB | 1.00 GiB | 1.00× |
| 32,768 | 2.00 GiB | 2.00 GiB | 1.00× |
| 65,536 | 4.00 GiB | 4.00 GiB | 1.00× |
| 131,072 | 8.00 GiB | 8.00 GiB | 1.00× |