// RUN — FITS

Can the GMKtec EVO-X3 run BERT-Large?

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

BERT-Large on the GMKtec EVO-X3

QuantFits?Max contextSpeedBeyond context
FP16 yes 1725k 244.7-297.4 t/s ~9.6-11.6 (slow)
Q8 yes 1730k 489.4-594.8 t/s ~19.1-23.2 (slow)
Q4 yes 1732k 873.9-1062.2 t/s ~34.1-41.5 (slow)

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

→ All models on the GMKtec EVO-X3 → BERT-Large specs → Other hardware

FAQ

Can the GMKtec EVO-X3 run BERT-Large?▶

Yes. At Q4 it generates ~873.9-1062.2 tok/s and fits up to 1732k context; beyond that it spills to SSD swap and slows to ~34.1-41.5 tok/s.

How much context fits?▶

At Q4, up to about 1732k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.