// RUN โ€” FITS

Can the Jetson Thor T3000 run Granite SWASH 3B?

Yes โ€” here's how fast and up to what context, with real measured numbers.

Granite SWASH 3B on the Jetson Thor T3000

QuantFits?Max contextSpeedBeyond context
FP16 yes 324k 33-39.3 t/s ~1.2-1.4 (slow)
Q8 yes 370k 66-78.6 t/s ~2.4-2.9 (slow)
Q4 yes 391k 117.8-140.4 t/s ~4.3-5.1 (slow)

Estimate (memory-bound). Beyond "max context" the KV cache spills to SSD swap and speed drops to the "beyond" figure.

FAQ

Can the Jetson Thor T3000 run Granite SWASH 3B?โ–ถ

Yes. At Q4 it generates ~117.8-140.4 tok/s and fits up to 391k context; beyond that it spills to SSD swap and slows to ~4.3-5.1 tok/s.

How much context fits?โ–ถ

At Q4, up to about 391k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.