🔌
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 Gemma 3 27B?
Yes — here's how fast and up to what context, with real measured numbers.
Gemma 3 27B on the NVIDIA DGX Spark
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 228k | 3.7-4.4 t/s | ~0.1-0.2 (slow) |
| Q8 | yes | 331k | 7.4-8.8 t/s | ~0.3-0.3 (slow) |
| Q4 | yes | 376k | 13.2-15.7 t/s | ~0.5-0.6 (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 Gemma 3 27B?▶
Yes. At Q4 it generates ~13.2-15.7 tok/s and fits up to 376k context; beyond that it spills to SSD swap and slows to ~0.5-0.6 tok/s.
How much context fits?▶
At Q4, up to about 376k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.