DeepSeek V4-Pro — 显存、速度与本地部署
作者： Jakub Rusinowski · 最后更新：
DeepSeek's flagship MoE — 1.6T total, 49B active, 1M context, MIT license. Not realistically self-hostable on any consumer or prosumer setup: even Q2 quantization is estimated around 400GB, requiring an 8x80GB+ H100/H800/B200-class server. Included for reference; the API or DeepSeek's hosted chat is the practical access path for almost everyone.
DeepSeek V4-Pro 在 Q2 (experimental, datacenter only) 下约需 967 GB 显存——量化权重加框架开销，不含 KV 缓存。在 Apple Silicon 上，这部分来自统一内存。
按量化级别的显存与速度
计算基准：NVIDIA RTX 4090 (24 GB)。仅含权重与开销：该模型架构未公开，因此未计入 KV 缓存。
| 量化 | 显存 | 显存 | 速度（估算） | 适配 |
|---|---|---|---|---|
| Q2_K 2.63 bpw | 526.8 GB | — | 放不下 | |
| Q3_K_M 3.41 bpw | 682.8 GB | — | 放不下 | |
| Q4_K_M 4.83 bpw | 966.8 GB | — | 放不下 | |
| Q5_K_M 5.67 bpw | 1134.8 GB | — | 放不下 | |
| Q6_K 6.56 bpw | 1312.8 GB | — | 放不下 | |
| Q8_0 8.50 bpw | 1700.8 GB | — | 放不下 | |
| F16 16.00 bpw | 3200.8 GB | — | 放不下 |
黑色标记 = NVIDIA RTX 4090 (24 GB) 上的可用显存。 估算来自内存带宽屋顶线模型，详见 方法说明页. DeepSeek V4-Pro 显存计算器 →
运行 DeepSeek V4-Pro
如何运行 DeepSeek V4-Pro
安装 Ollama，然后运行：
ollama run deepseek-v4规格
Verified — Checked against the primary source — the model card or the vendor spec page — and corroborated by a second independent source.
- 参数量
- 1.6 Trillion (49B active)
- 上下文窗口
- 1,000,000
- 架构
- Mixture-of-Experts
- 提供商
- DeepSeek
- 许可证
- MIT
- 规格量化
- Q2 (experimental, datacenter only)
- 系统内存
- 1024 GB
- 记录更新于
- 2026-04-24
Commercial use permitted. No usage restrictions beyond attribution.
质量与使用场景
评分由模型作者或独立评测方发布——衡量质量而非吞吐量，并非我们实测。
我的 GPU 能运行 DeepSeek V4-Pro 吗？
- DeepSeek V4-Pro 在 AMD Ryzen AI Max+ 395 上
- DeepSeek V4-Pro 在 Apple M1 Ultra 上
- DeepSeek V4-Pro 在 Apple M2 Max 上
- DeepSeek V4-Pro 在 Apple M2 Ultra 上
- DeepSeek V4-Pro 在 Apple M4 Max 上
- DeepSeek V4-Pro 在 Apple M5 Max 上
- DeepSeek V4-Pro 在 NVIDIA A100 80GB (PCIe) 上
- DeepSeek V4-Pro 在 NVIDIA DGX Spark 上
- DeepSeek V4-Pro 在 NVIDIA H100 80GB (PCIe) 上
- DeepSeek V4-Pro 在 NVIDIA RTX PRO 6000 Blackwell 上
DeepSeek V4 的其他尺寸
DeepSeek V4-Pro — 常见问题
How much VRAM does DeepSeek V4-Pro need?
About 967 GB at Q2 (experimental, datacenter only) — quantized weights plus framework overhead, before any KV cache. The cache grows with context length and is added on top; the table above folds it in. Apple Silicon counts unified memory toward the same figure.
Does DeepSeek V4-Pro run on an RTX 4090 (24 GB)?
No. DeepSeek V4-Pro needs about 967 GB at Q2 (experimental, datacenter only), more than a single RTX 4090's 24 GB. It needs a larger card, several GPUs, or Apple Silicon with enough unified memory — or it runs with part of the weights offloaded to system RAM, which is much slower.
How do I run DeepSeek V4-Pro locally?
Install Ollama and run `ollama run deepseek-v4`. That pulls the weights and starts a local OpenAI-compatible endpoint; after the download nothing leaves the machine.
What other sizes does DeepSeek V4 come in?
DeepSeek V4-Flash (172 GB), DeepSeek V4-Pro (967 GB). Every size shares the family's training and licence; the larger ones score higher and need proportionally more memory.