DeepSeek-V4-Flash
by deepseek-ai
158B params · text-generation · 1.5k likes · 2.4M downloads
DeepSeek-V4-Flash is a 158B parameter model. At Q4 quantization it requires 79GB of VRAM. It runs comfortably on H200 SXM 141 GB (44 tok/s).
Inference providers
| Provider | $/1M in | $/1M out | Throughput |
|---|---|---|---|
| Novita | 86 tok/s | ||
| Fireworks | 59 tok/s | ||
| Featherless | |||
| deepinfra | 13 tok/s |
GPU compatibility
| GPU | VRAM | Q4 Decode | Verdict |
|---|---|---|---|
| M4 Max 128GB | 128GB | 3 tok/s | tight |
| H200 SXM 141 GB | 141GB | 44 tok/s | comfortable |
| M2 Ultra 192GB | 192GB | 4 tok/s | tight |