FHE Benchmarking Results - ML Inference (Medium Batch - 1000)

Submitter					Bandwidth			Quality	Timing (harness)									Timing (server)
Name	Date	Env	R/L	Model	Keys	Input	Result	Accuracy	Total	DB Gen	Keygen	Model Prep	Input Gen	Input Prep	Input Enc	Compute	Decrypt	Total	DB Gen	Keygen	Model Prep	Input Gen	Input Prep	Input Enc	Compute	Decrypt	I/O
Lattica-ai	2026‑02‑23 16:49:29	GPU	R	mlp	16.5M	196M	2.5M	0.972	32.26s	7.3829s	1.8578s	200us	3.1782s	1.8901s	5.1136s	3.5153s	2.2483s								605.3ms
Lattica-ai	2026‑05‑25 08:42:44	GPU	R	mlp	7.3M	98M	1.3M	0.972	28.79s	7.0801s	1.7128s	200us	3.1457s	1.8944s	3.1177s	2.5179s	1.9247s								214.7ms
Reference	2026‑02‑20 03:38:16	CPU	L	mlp	1G	4.9G	1001.6M	0.974	1.8286h	4.2678s	966ms	3.9ms	2.0617s	186ms	35.93s	1.8091h	24.74s	1.7744h							1.7744h