DeepSeek · 158B (13B active) · 專家混合（MoE）
Efficient long-context V4 — 13B active, 1M context 看看你的 GPU 或 Mac 跑不跑得動 DeepSeek V4 Flash——最低 88.3 GB，建議 147.1 GB。
2026-041024K context
專家混合（MoE）
專家總数： 256
啟用專家： 6
啟用参数： 13.0B
| 量化 | 位元 | VRAM | 品質 | 状態 |
|---|---|---|---|---|
| Q2_K | 2 | 51.1 GB | low | — |
| Q3_K_M | 3 | 71.3 GB | moderate | — |
| Q4_K_M | 4 | 81.4 GB | good | — |
| Q5_K_M | 5 | 101.7 GB | good | — |
| Q6_K | 6 | 121.9 GB | excellent | — |
| Q8_0 | 8 | 162.4 GB | excellent | — |
| F16 | 16 | 324.2 GB | lossless | — |
Can I run DeepSeek V4 Flash locally?
- Can I run DeepSeek V4 Flash locally?
- DeepSeek V4 Flash needs about 88.3 GB of memory at a minimum and 147.1 GB recommended. Open this page to grade it against your GPU or Mac, then run it with runai, Ollama or LM Studio.
- How much VRAM does DeepSeek V4 Flash need?
- At Q4_K_M, DeepSeek V4 Flash uses about 81.4 GB of VRAM. Higher quants need more memory; lower quants fit tighter cards with a quality tradeoff.