🔌
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 Japanese-LM-3.6B?
Yes — here's how fast and up to what context, with real measured numbers.
Japanese-LM-3.6B on the NVIDIA DGX Spark
| Quant | Fits? | Max context | Speed | Beyond context |
|---|---|---|---|---|
| FP16 | yes | 1625k | 27.7-33 t/s | ~1-1.2 (slow) |
| Q8 | yes | 1680k | 55.4-66 t/s | ~2-2.4 (slow) |
| Q4 | yes | 1704k | 98.9-117.8 t/s | ~3.6-4.3 (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 Japanese-LM-3.6B?▶
Yes. At Q4 it generates ~98.9-117.8 tok/s and fits up to 1704k context; beyond that it spills to SSD swap and slows to ~3.6-4.3 tok/s.
How much context fits?▶
At Q4, up to about 1704k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.