community pool / model answer
DeepSeek-V4-Flash-W4A16-FP8
chat · pastapaul/DeepSeek-V4-Flash-W4A16-FP8
11.9 tok/s · model median · 1 runs in the model record.
| Reference rig | Quantization | Median result | Basis | TTFT (ms) | Peak VRAM | Max context | Engines |
|---|---|---|---|---|---|---|---|
| GB10 Grace Blackwell 256GB | 4-bit | 11.9 tok/s | reported | 597 | - GB | 1,048,576 | vllm |
Context tested
The table above is the tested context for each exact rig and quantization cell. Empty fields remain empty rather than being filled with an estimate.