Can I Run Llama 3.2 Family on NVIDIA DGX Spark?

作者： Jakub Rusinowski · 最后更新： 2026年7月10日

Yes — the Llama 3.2 Family family runs on the NVIDIA DGX Spark.

联盟营销声明: 本页部分链接为联盟推广链接——如果你通过它们购买，LLM Configurator 可能会获得佣金，而你无需支付任何额外费用。作为亚马逊联盟成员（Amazon Associate），LLM Configurator 会从符合条件的购买中获得收益。

在亚马逊查看价格 — NVIDIA DGX Spark (128GB)

NVIDIA DGX Spark Specs

VRAM128 GB
Memory Bandwidth273 GB/s

Llama 3.2 Family Variants That Fit

Llama 3.2 90B Vision InstructQ4_K_M · 54 GB · ~5 tok/s (est.)
Llama 3.2 11B Vision InstructQ4_K_M · 7.8 GB · ~35 tok/s (est.)
Llama 3.2 3B InstructQ4_K_M · 2.2 GB · ~124 tok/s (est.)
Llama 3.2 1B InstructQ4_K_M · 0.8 GB · ~341 tok/s (est.)

FAQ

Can I run Llama 3.2 Family on the NVIDIA DGX Spark?

Yes — the Llama 3.2 Family family runs on the NVIDIA DGX Spark.

Which Llama 3.2 Family variant fits best on the NVIDIA DGX Spark?

Llama 3.2 90B Vision Instruct at Q4_K_M quantization (54 GB), estimated ~5 tokens/sec.

Llama 3.2 Family on Other GPUs

Popular Models on the NVIDIA DGX Spark

VRAM Tier

Troubleshooting

Buying Guide

← Can I Run It? | Llama 3.2 Family Model Page | NVIDIA DGX Spark GPU Page | Check Your Hardware