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

Processor	AI Accelerator	Year	Lib	Inference Mode	Inception-V3		DeepLabV3+		SRGAN		U-Net		VSR		Average FPS / Watt
					FPS	FPS/Watt	FPS	FPS/Watt	FPS	FPS/Watt	FPS	FPS/Watt	FPS	FPS/Watt
Snapdragon 888	DSP (Hexagon 780) + GPU (Adreno 660)	2020	qh.qg	SUSTAINED SPEED	481	261	77.7	51.1	41.8	20	62.6	41	254	124	67.1
Snapdragon 8 Elite	Hexagon DSP / HTP Gen 4	2024	qh.qh	FAST SINGLE ANSWER	2089	232	600	76.9	223	14.6	259	32.3	779	99.7	60.9
Snapdragon 8 Gen 3	Hexagon DSP / HTP Gen 3	2023	qh.qh	FAST SINGLE ANSWER	1728	242	484	79.1	147	13.8	210	32.8	503	78.7	58.4
Snapdragon 8 Gen 2	Hexagon DSP / HTP Gen 2	2022	qh.qh	SUSTAINED SPEED	1295	229	303	74.2	120	13.9	146	32.4	351	72.3	56.1
Snapdragon 8 Gen 1	Hexagon DSP / HTP	2021	qh.qh	SUSTAINED SPEED	663	145	131	36.8	63.6	12.3	102	21.9	309	51	37.4
Dimensity 9400	NPU 890	2025	mm	FAST SINGLE ANSWER	908	98.9	256	31	131	13.5	293	36.5	310	44.4	36.8
Dimensity 8400	NPU 880	2024	mm	FAST SINGLE ANSWER	469	91.7	134	29.8	81.7	13.6	164	35.2	253	50	36.6
Dimensity 8300 Ultra	APU 780	2023	mm	FAST SINGLE ANSWER	545	105	139	31.8	78.1	10.5	153	36.1	204	44.6	35.5
Dimensity 9300	APU 790	2023	mm	FAST SINGLE ANSWER	756	108	206	30.3	138	13.8	237	33	206	32.4	34.4
Snapdragon 7 Gen 1	Hexagon DSP / HTP	2022	qh.qh	FAST SINGLE ANSWER	206	96.7	34.6	17.5	19.3	7.87	36.6	18.8	134	80	28.9
Dimensity 9200	APU 690	2022	mm	SUSTAINED SPEED	156	95.8	76.5	21.3	39.9	12.9	65.3	18.1	55.6	19.5	24.8
Dimensity 9000	APU 590	2021	mm	SUSTAINED SPEED	130	93.6	61.5	17.6	27.9	13.1	51.5	14.6	43.6	14.6	21.5
Dimensity 800	APU 3.0 (4 cores)	2020	nn	SUSTAINED SPEED	44.9	96.2	13.8	15.5	3.59	6.07	8.92	10.9	5.76	9.9	15.8
Dimensity 820	APU 3.0 (4 cores)	2020	nn	SUSTAINED SPEED	41.8	97.8	10.8	14.2	2.7	5.35	5.73	9.49	5.31	9.73	14.7
Google Tensor G2	Google Tensor TPU 2.0	2022	nn	SUSTAINED SPEED	282	80	80.1	21.3	8.74	2.08	34.5	8.48	28	10.5	12.6
Dimensity 1000+	APU 3.0 (6 cores)	2019	nn	SUSTAINED SPEED	54.6	89.6	14.1	15.9	4.53	6.53	8.49	10.2	1.61	2.76	12.1
Google Tensor	Google Tensor TPU	2021	nn	SUSTAINED SPEED	259	81.8	33.6	9.44	3.89	0.74	18.7	6.23	6.34	4.65	7.0
Snapdragon 865	DSP (Hexagon 698) + GPU (Adreno 650)	2019	hg	SUSTAINED SPEED	93.9	78.5	7.71	4.72	5.94	1.99	8.54	3.43	4.96	3.28	6.1
Snapdragon 855	DSP (Hexagon 690) + GPU (Adreno 640)	2018	hg	SUSTAINED SPEED	84.7	53.9	7.37	4.46	5.44	1.42	7.93	3.1	3.93	2.76	4.9
Snapdragon 845	DSP (Hexagon 685) + GPU (Adreno 630)	2018	hg	SUSTAINED SPEED	29.2	38.4	3.89	3.09	2.34	1.45	2.9	1.82	2	2.05	3.6
Snapdragon 710	DSP (Hexagon 685) + GPU (Adreno 616)	2018	hg	SUSTAINED SPEED	23.1	23	3.61	3.18	2.41	1.49	2.16	1.67	2.02	2.26	3.3
Exynos 2100	GPU (Mali-G78 MP14)	2021	gg	SUSTAINED SPEED	28.7	5.84	12.1	2.01	2.69	1.05	8.6	1.46	7.51	5.03	2.5
Kirin 990 5G	GPU (Mali-G76 MP10)	2019	gg	SUSTAINED SPEED	18	7.95	7.54	2.38	3.03	0.8	5.18	1.66	9.69	3.65	2.5
Kirin 9000	GPU (Mali-G78 MP24)	2020	gg	SUSTAINED SPEED	23	8.16	13.2	2.29	4.7	0.67	10.1	1.52	13.7	3.84	2.4
Exynos 9820 Octa	GPU (Mali-G76 MP12)	2018	gg	SUSTAINED SPEED	16.6	6.49	6.64	1.75	2.96	0.68	4.67	1.16	9.48	2.29	1.8
Kirin 980	GPU (Mali-G76 MP10)	2018	gg	SUSTAINED SPEED	15	5.68	5.88	1.82	2.45	0.56	4.06	1.23	7.3	2.71	1.8

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

n   -   Android NNAPI

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

h     -   Qualcomm Hexagon NN Direct Delegate

qh   -   Qualcomm QNN HTP Delegate

qd   -   Qualcomm QNN DSP Delegate

qg   -   Qualcomm QNN GPU Delegate

e   -   Samsung ENN Delegate

m  -   MediaTek Neuron Delegate

i    -    Apple CoreML 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