// RUN โ€” FITS

Can the NVIDIA GeForce RTX 3090 run Surya?

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

Surya on the NVIDIA GeForce RTX 3090

QuantFits?Max contextSpeedBeyond context
FP16 yes 303k 933.4-1112.5 t/s ~39.9-47.5 (slow)
Q8 yes 308k 1866.9-2224.9 t/s ~79.8-95.1 (slow)
Q4 yes 311k 3333.7-3973.1 t/s ~142.5-169.8 (slow)

Estimate (memory-bound). Beyond "max context" the KV cache spills to system RAM offload (depends on your system RAM) and speed drops to the "beyond" figure.

FAQ

Can the NVIDIA GeForce RTX 3090 run Surya?โ–ถ

Yes. At Q4 it generates ~3333.7-3973.1 tok/s and fits up to 311k context; beyond that it spills to system RAM offload (depends on your system RAM) and slows to ~142.5-169.8 tok/s.

How much context fits?โ–ถ

At Q4, up to about 311k tokens stay in fast memory; longer context spills to system RAM offload (depends on your system RAM) and throughput collapses.