• BENCHMARK
• NEWS
• RANKING
• AI-TESTS
• RESEARCH  
• BURNOUT

Processor	CPU Cores	AI Accelerator	Year	Lib	CPU-Q Score	CPU-F Score	INT8 NNAPI 1.1	INT8 NNAPI 1.2	INT8 Accuracy	FP16 NNAPI 1.1	FP16 NNAPI 1.2	FP16 Accuracy	INT8 Parallel	FP16 Parallel	LSTM	AI Score, K
Synaptics VideoSmart VS680-A0	4x1.8 GHz Cortex-A73	NPU (Vivante VIP9000, 7 TOPS)	2021	nc	752	2316	14763	11101	94.4	151	248	59.5	420	43	186	28.7
Qualcomm QCS605	2x2.5 GHz Kryo 300/G & 6x1.7 GHz Kryo 300/S	DSP (Hex. 685) + GPU (Adreno 615)	2018	hg	1099	3487	5846	4501	90.3	4542	6847	54.6	755	752	1232	25.7
i.MX 8M Plus	4x2.0 GHz Cortex-A53	NPU (Vivante VIP8000)	2020	nc	488	1469	7072	956	94.4	177	217	60.9	320	50	9	10.3
Synaptics VS640	4x2.0 GHz Cortex-A55	NPU (Vivante VIP9000, 1 TOPS)	2021	nc	295	1206	5410	1208	94.4	348	849	100	405	47	530	10.0
MediaTek i500	4x2.0 GHz Cortex-A73 & 4x2.0 GHz Cortex-A53	APU + GPU (Mali-G72 MP3)	2019	nn	871	2607	906	489	100	1207	1352	42.2	150	240	27	6.6
Amlogic A311D	4x2.2 GHz Cortex-A73 & 2x1.8 GHz Cortex-A55	none	2018	cc	802	2442	189	289	98.8	460	799	100	15	50	687	5.8
Amlogic S922X	4x1.8 GHz Cortex-A73 & 2x1.9 GHz Cortex-A53	none	2018	cc	778	2327	188	287	98.8	441	773	100	15	48	637	5.6
Amlogic S922X	4x1.8 GHz Cortex-A73 & 2x1.9 GHz Cortex-A53	GPU (Mali-G52 MP4)	2018	cg	779	2311	307	288	98.8	564	1064	85.6	16	38	44	5.3
Broadcom BCM2711	4x1.5 GHz Cortex-A72	none	2019	cc	645	1946	424	377	97	364	641	100	36	68	563	5.1
Rockchip RK3566	4x1.8 GHz Cortex-A55	none	2020	cc	942	1327	593	508	96.8	351	594	100	44	61	423	4.9
JLQ JA310	4x1.5 GHz Cortex-A55	none	2020	cc	855	1436	525	485	96.8	344	624	100	40	57	422	4.8
Rockchip RK3288	4x1.8 GHz Cortex-A17	none	2014	cc	417	1075	297	262	100	370	723	100	25	32	494	3.8
Amlogic S905D	4x2.0 GHz Cortex-A53	none	2016	cc	285	1033	223	190	100	422	765	100	20	37	398	3.5
Amlogic S905X2	4x1.8 GHz Cortex-A53	none	2018	cc	329	1257	237	203	98.8	226	398	100	24	30	272	3.0
Amlogic S905X3	4x1.9 GHz Cortex-A55	none	2019	cc	342	1246	272	229	98.8	217	372	100	25	30	238	3.0
Amlogic S905X4	4x2.0 GHz Cortex-A53	none	2019	cc	314	1155	235	206	97	224	397	100	23	45	337	3.0
Amlogic S905X	4x1.5 GHz Cortex-A53	none	2016	cc	211	776	163	132	100	290	509	100	17	29	340	2.6
Amlogic S905W	4x1.5 GHz Cortex-A53	none	2017	cc	218	756	163	144	100	282	499	100	18	25	326	2.5
Rockchip RK3399	2x1.8 GHz Cortex-A72 & 4x1.4 GHz Cortex-A55	none	2016	cc	404	958	73	121	98.8	220	377	100	5	34	282	2.5
Amlogic S905	4x1.5 GHz Cortex-A53	none	2015	cc	209	765	133	130	100	241	521	100	17	29	283	2.4
Amlogic S912	4x1.5 GHz Cortex-A53 & 4x1.0 GHz Cortex-A53	none	2016	cc	273	882	214	187	98.8	174	332	100	19	20	208	2.3
Rockchip RK3318	4x1.5 GHz Cortex-A53	none	2019	cc	278	780	187	170	96.8	167	306	100	21	32	226	2.2
Synaptics VideoSmart BG5CT	4x1.6 GHz Cortex-A53	none	2017	cc	292	1136	222	166	97	112	164	86.2	10	6	17	2.1
Allwinner H616	4x2.0 GHz Cortex-A53	none	2020	cc	225	714	180	151	97	145	262	100	16	28	225	2.0
Rockchip PX30	4x1.5 GHz Cortex-A35	none	2018	cc	150	425	113	98	97	106	182	100	14	23	190	1.3
Allwinner T8	8x1.6 GHz Cortex-A7	none	2016	cc	138	332	99	78	100	140	215	100	20	27	136	1.2
AutoChips AC8227L	4x1.1 GHz Cortex-A7	none	2018	cc	122	322	90	78	100	111	209	100	10	12	146	1.1
Allwinner T3	4x1.2 GHz Cortex-A7	none	2016	cc	119	314	83	62	100	120	187	100	12	15	126	1.1
NXP i.MX 8QuadXPlus	4x1.2 GHz Cortex-A35	none	2017	cc	197	455	144	65	95.3	78	85	85.2	15	14	35	1.1

xy  -  This device accelerates quantized [x] and floating-point [y] models using the:

n   -   Android NNAPI

g   -   TensorFlow Lite GPU Delegate  (OpenCL / OpenGL based)

h   -   Qualcomm Hexagon NN Direct Delegate

e   -   Samsung Eden Delegate

m  -   MediaTek Neuron Delegate

c   -   TensorFlow Lite / NNAPI default CPU backend

1  -  This SoC might be using unofficial / prototype hardware or drivers

2  -  These are the results of an early prototype. The results of the commercial SoC might be different