community pool / model answer
DeepSeek-V4-Flash-DSpark
chat · deepseek-ai/DeepSeek-V4-Flash-DSpark
212.9 tok/s · model median · 4 runs in the model record.
| Reference rig | Quantization | Median result | Basis | TTFT (ms) | Peak VRAM | Max context | Engines |
|---|---|---|---|---|---|---|---|
| GB10 Grace Blackwell 128GB | 4-bit | 46.4 tok/s | reported | 175 | - GB | 1,048,576 | vllm |
| RTX PRO 6000 Blackwell 96GB ×2 | 8-bit | 261.8 tok/s | measured | 4,830 | 184 GB | 262,144 | 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.