Nemotron 3.5 Lightning 30B-A3B — 显存、速度与本地部署
作者： Jakub Rusinowski · 最后更新：
30B total, 3B active, pre-trained on over 20T tokens and released with speculative-decoding companions. Hybrid Mamba-2 + MoE + attention, so its cache does not grow the way a pure transformer's would. BF16 and NVFP4 builds are published.
Nemotron 3.5 Lightning 30B-A3B 在 Q4_K_M 下约需 19 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 | 10.7 GB | ~212 tok/s | 可运行 | |
| Q3_K_M 3.41 bpw | 13.6 GB | ~196 tok/s | 可运行 | |
| Q4_K_M 4.83 bpw | 18.9 GB | ~172 tok/s | 可运行 | |
| Q5_K_M 5.67 bpw | 22.1 GB | ~160 tok/s | 勉强 | |
| Q6_K 6.56 bpw | 25.4 GB | ~22 tok/s | 需卸载 | |
| Q8_0 8.50 bpw | 32.7 GB | ~20 tok/s | 需卸载 | |
| F16 16.00 bpw | 60.8 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. Nemotron 3.5 Lightning 30B-A3B 显存计算器 →
运行 Nemotron 3.5 Lightning 30B-A3B
目录中能运行 Nemotron 3.5 Lightning 30B-A3B 的最便宜 GPU 是 AMD Radeon RX 7900 XTX (24 GB).
或在 Vast.ai 比较，低至 $0.35/小时 (typical low · varies)
作为亚马逊联盟成员，我们从符合条件的购买中获得收入。云 GPU 链接为推荐链接——我们可能获得佣金，您无需额外付费。
如何运行 Nemotron 3.5 Lightning 30B-A3B
安装 Ollama，然后运行：
ollama run nemotron-3.5-lightning:30b规格
Corroborated — Two or more independent sources agree on these figures, but the model card itself was not retrieved. Treat the numbers as good rather than confirmed. 尚未确认： arch.kv_layers.
- 参数量
- 30 Billion (3B active)
- 上下文窗口
- 1,000,000
- 架构
- Hybrid Mamba-2 + MoE
- 提供商
- NVIDIA
- 许可证
- OpenMDW-1.1
- Practical deployment
- Single consumer GPU
- 规格量化
- Q4_K_M
- 系统内存
- 32 GB
- 记录更新于
- 2026-09-19
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 Nemotron 3.5 Lightning 30B-A3B 吗？
Nemotron 3.5 的其他尺寸
Nemotron 3.5 Lightning 30B-A3B — 常见问题
How much VRAM does Nemotron 3.5 Lightning 30B-A3B need?
About 19 GB at Q4_K_M — 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 Nemotron 3.5 Lightning 30B-A3B run on an RTX 4090 (24 GB)?
Yes. Nemotron 3.5 Lightning 30B-A3B needs about 19 GB at Q4_K_M, inside a 24 GB card, at an estimated 172 tokens/sec.
How do I run Nemotron 3.5 Lightning 30B-A3B locally?
Install Ollama and run `ollama run nemotron-3.5-lightning:30b`. That pulls the weights and starts a local OpenAI-compatible endpoint; after the download nothing leaves the machine.
What other sizes does Nemotron 3.5 come in?
Nemotron 3.5 Lightning 30B-A3B (19 GB), Nemotron 3.5 Bolt 74B (23 GB). Every size shares the family's training and licence; the larger ones score higher and need proportionally more memory.