How much memory does Qwen 3.6 27B need in NVFP4?
Qwen 3.6 27B has 27 billion parameters. The NVFP4 format decides how much memory is needed to load them.
How is this footprint calculated?
Qwen 3.6 27B totals 27 billion parameters. The NVFP4 format takes 0.5 byte per parameter. The product gives the weights, to which a 15% runtime margin is added for activations and buffers. The attention cache is not included: it grows with context and concurrency.
What does the format change?
NVFP4 brings Blackwell format, native FP4 compute.
| Format | Weights in memory | Compatible platforms |
|---|---|---|
| FP16 | 62.1 GB | 8 |
| FP8 | 31.0 GB | 8 |
| NVFP4 | 15.5 GB | 8 |
| Q4 | 15.5 GB | 8 |
Which platforms qualify?
- DGX Spark, 128 GB
- Mac Studio Ultra, 512 GB
- DGX Station, 748 GB
- RTX PRO 6000 server, 768 GB
- H200 SXM server, 1,128 GB
- B200 SXM, 1,440 GB
See the Qwen 3.6 27B fact sheet.
How much memory for Qwen 3.6 27B in NVFP4?
About 15.5 GB for the weights, excluding the attention cache.
Which format should I choose for Qwen 3.6 27B?
NVFP4 brings Blackwell format, native FP4 compute. The most precise format that fits the target machine remains the best choice.
Method
Parameter counts and memory figures come from the site fact sheets. Memory footprints are calculated, not measured: a reading on real hardware may differ depending on the engine and the exact weight format. Prices are indicative and not contractual.