🔌
New: connect Claude & other AIs to GenAIList over MCP — research the catalog and contribute to the shared knowledge base. Learn how →
// RUN — FITS
Can the NVIDIA DGX Spark run Llama 3.2 90B?
Yes — here's how fast and up to what context, with real measured numbers.
Llama 3.2 90B on the NVIDIA DGX Spark
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | no | — | — | — |
| Q8 | yes | 77k | 2.2-2.7 t/s | ~0.1-0.1 (slow) |
| Q4 | yes | 196k | 4-4.8 t/s | ~0.1-0.2 (slow) |
Estimate (memory-bound). Beyond "max context" the KV cache spills to SSD swap and speed drops to the "beyond" figure.
FAQ
Can the NVIDIA DGX Spark run Llama 3.2 90B?▶
Yes. At Q4 it generates ~4-4.8 tok/s and fits up to 196k context; beyond that it spills to SSD swap and slows to ~0.1-0.2 tok/s.
How much context fits?▶
At Q4, up to about 196k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.