GPU Silicon Database: NVIDIA GPUs by Architecture

Every discrete desktop, workstation and datacenter graphics chip from NVIDIA, AMD and Intel since unified shaders arrived in 2006, grouped by architecture: the die underneath each card, its transistor count and size, process node, shader count, clocks, compute, memory, board power and launch price. 809 GPUs, 127 of them benchmarked here.

NVIDIA 459 AMD 334 Intel 16

Tesla 2006–2010 Tesla 2.0 2007–2012 Fermi 2010–2013 Kepler 2012–2014 Maxwell 2014–2016 Pascal 2016–2019 Volta 2017–2018 Turing 2018–2022 Ampere 2020–2024 Hopper 2022–2024 Ada Lovelace 2022–2024 Blackwell 2024–2025

Tesla

2006–2010 · TSMC 90 nm / TSMC 80 nm / UMC 65 nm / TSMC 55 nm / TSMC 40 nm · 82 GPUs · up to 754M transistors

NVIDIA's first unified-shader architecture launched with the GeForce 8800 GTX in November 2006. Separate vertex and pixel pipelines gave way to 128 scalar stream processors that could run either job, built for DirectX 10, and CUDA shipped alongside it in 2007. The 65 nm G92 refresh carried the design through the GeForce 8800 GT and the GeForce 9 series.

GPU	Launch	Die	Node	Transistors	Die size	Shaders	Clock	FP32 TFLOPS	Memory	Bandwidth	TDP	Launch price
Desktop 62
GeForce 8800 GTS 640	November 8, 2006	G80-100-K0-A2	TSMC 9 nm	681M	484 mm²1.4 M/mm²	96	513MHz	0.342	640MB GDDR3320-bit	63.4GB/s	143 W	$449
GeForce 8800 GTX	November 8, 2006	G80-300-A2	TSMC 9 nm	681M	484 mm²1.4 M/mm²	128	576MHz	0.518	768MB GDDR3384-bit	86.4GB/s	155 W	$599
GeForce 8800 GTS 320	February 12, 2007	G80-100-K0-A2	TSMC 9 nm	681M	484 mm²1.4 M/mm²	96	513MHz	0.342	320MB GDDR3320-bit	63.4GB/s	143 W	$269
GeForce 8300 GSOEM	April 17, 2007	G86S	TSMC 8 nm	210M	127 mm²1.7 M/mm²	8	459MHz	0.015	128MB DDR264-bit	6.4GB/s	40 W	–
GeForce 8400 GSOEM	April 17, 2007	G86-213-A2	TSMC 4 nm	210M	127 mm²1.7 M/mm²	16	459MHz	0.029	256MB / 512MB DDR264-bit	6.4GB/s	40 W	–
GeForce 8500 GT	April 17, 2007	G86-300-A2	TSMC 8 nm	210M	127 mm²1.7 M/mm²	16	459MHz	0.029	256MB / 512MB / 1GB GDDR3128-bit	12.8GB/s	30 W	$129
GeForce 8600 GSOEM	April 17, 2007	G84	TSMC 8 nm	289M	169 mm²1.7 M/mm²	16	540MHz	0.038	512MB DDR2128-bit	12.8GB/s	47 W	–
GeForce 8600 GT	April 17, 2007	G84-303-A2	TSMC 8 nm	289M	169 mm²1.7 M/mm²	32	540MHz	0.076	512MB / 1GB GDDR3128-bit	22.4GB/s	47 W	$159
GeForce 8600 GT Mac Edition	April 17, 2007	G84-303-A2	TSMC 8 nm	289M	169 mm²1.7 M/mm²	32	600MHz	0.093	128MB GDDR3128-bit	24GB/s	47 W	$159
GeForce 8600 GTS	April 17, 2007	G84-400-A2	TSMC 8 nm	289M	169 mm²1.7 M/mm²	32	675MHz	0.093	256MB / 512MB GDDR3128-bit	32GB/s	60 W	$199
GeForce 8800 Ultra	May 2, 2007	G80-450-A3	TSMC 9 nm	681M	484 mm²1.4 M/mm²	128	612MHz	0.581	768MB GDDR3384-bit	103.7GB/s	171 W	$829
GeForce 8600 GTS Mac Edition	September 21, 2007	G84-400-A2	TSMC 8 nm	289M	169 mm²1.7 M/mm²	32	675MHz	0.093	512MB GDDR3128-bit	32.3GB/s	75 W	$199
GeForce 8800 GT	October 29, 2007	G92-270-A2	TSMC 65 nm	754M	324 mm²2.3 M/mm²	112	600MHz	0.336	512MB GDDR3256-bit	57.6GB/s	125 W	$349
GeForce 8800 GTS Core 112	November 19, 2007	G80-100-K0-A2	TSMC 9 nm	681M	484 mm²1.4 M/mm²	112	500MHz	0.403	640MB GDDR3320-bit	64GB/s	150 W	$399
GeForce 8400OEM	December 4, 2007	G98S	UMC 65 nm	210M	86 mm²2.4 M/mm²	8	540MHz	0.021	256MB DDR264-bit	8GB/s	25 W	–
GeForce 8400 GS PCI	December 4, 2007	G98S	UMC 65 nm	210M	86 mm²2.4 M/mm²	8	567MHz	0.022	512MB DDR264-bit	5.3GB/s	25 W	$80
GeForce 8400 GS Rev. 2	December 4, 2007	G98-409-U2	UMC 65 nm	210M	86 mm²2.4 M/mm²	8	567MHz	0.022	512MB DDR264-bit	6.4GB/s	25 W	$60
GeForce 8800 GTS 512	December 11, 2007	G92-400-A2	TSMC 65 nm	754M	324 mm²2.3 M/mm²	128	650MHz	0.416	512MB GDDR3256-bit	52.5GB/s	135 W	$349
GeForce 8800 GSOEM	January 31, 2008	G92-150-A2	TSMC 65 nm	754M	324 mm²2.3 M/mm²	96	550MHz	0.264	384MB / 768MB GDDR3192-bit	38.4GB/s	105 W	–
GeForce 8800 GT Mac Edition	February 1, 2008	G92-270-A2	TSMC 65 nm	754M	324 mm²2.3 M/mm²	112	600MHz	0.336	512MB GDDR3256-bit	57.6GB/s	125 W	$349
GeForce 9600 GT	February 21, 2008	G94-300-A1	TSMC 65 nm	505M	240 mm²2.1 M/mm²	64	650MHz	0.208	512MB / 1GB GDDR3256-bit	57.6GB/s	95 W	$179
GeForce 9600 GT Green Edition	February 21, 2008	G94-300-A1 G94B	TSMC 65 nm	505M	240 mm²2.1 M/mm²	64	600MHz	0.192	512MB / 1GB GDDR3256-bit	44.8GB/s	59 W	$179
GeForce 9600 GT Rev. 2	February 21, 2008	G94-358-B1	TSMC 55 nm	505M	196 mm²2.6 M/mm²	64	650MHz	0.208	512MB / 1GB GDDR3256-bit	57.6GB/s	95 W	$179
GeForce 9800 GX22 GPUs	March 18, 2008	G92-450-A2	TSMC 65 nm	754M	324 mm²2.3 M/mm²	128	600MHz	0.384	512MB GDDR3256-bit	64GB/s	197 W	$599
GeForce 9800 GTX	March 28, 2008	G92-420-A2	TSMC 65 nm	754M	324 mm²2.3 M/mm²	128	675MHz	0.432	512MB GDDR3256-bit	70.4GB/s	140 W	$299
GeForce 9600 GSO	April 28, 2008	G92-150-A2	TSMC 65 nm	754M	324 mm²2.3 M/mm²	96	550MHz	0.264	384MB / 768MB / 1.5GB GDDR3192-bit	38.4GB/s	84 W	$169
GeForce 9300 GEOEM	June 1, 2008	G98	TSMC 55 nm	210M	80 mm²2.6 M/mm²	8	540MHz	0.021	256MB DDR264-bit	6.4GB/s	50 W	–
GeForce 9300 GSOEM	June 1, 2008	G98S	TSMC 65 nm	210M	86 mm²2.4 M/mm²	8	567MHz	0.022	512MB DDR264-bit	5.3GB/s	35 W	–
GeForce 9300 SEOEM	June 1, 2008	G98S	TSMC 65 nm	210M	86 mm²2.4 M/mm²	8	540MHz	0.021	256MB DDR264-bit	6.4GB/s	25 W	–
GeForce 9800 GTX+	July 16, 2008	G92-420-B1	TSMC 55 nm	754M	260 mm²2.9 M/mm²	128	738MHz	0.470	512MB GDDR3256-bit	70.4GB/s	141 W	$229
GeForce 9800 GT	July 21, 2008	G92-270-A2 G92-280-B1	TSMC 65 nm	754M	324 mm²2.3 M/mm²	112	600MHz	0.336	512MB / 1GB GDDR3256-bit	57GB/s	125 W	$159
GeForce 9800 GT Rebrand	July 21, 2008	G92-270-A2	TSMC 65 nm	754M	324 mm²2.3 M/mm²	128	725MHz	0.448	512MB GDDR3256-bit	64GB/s	125 W	$159
GeForce 9500 GSOEM	July 29, 2008	G96-259-A1	UMC 65 nm	314M	144 mm²2.2 M/mm²	32	550MHz	0.088	512MB DDR2128-bit	16.1GB/s	40 W	–
GeForce 9500 GS Rev. 2OEM	July 29, 2008	G96C	TSMC 55 nm	314M	121 mm²2.6 M/mm²	32	500MHz	0.080	512MB DDR2128-bit	16GB/s	40 W	–
GeForce 9500 GT	July 29, 2008	G96-300-C1	UMC 65 nm	314M	144 mm²2.2 M/mm²	32	600MHz	0.096	512MB GDDR3128-bit	32GB/s	50 W	–
GeForce 9500 GT Rev. 2	July 29, 2008	G96-309-B1	UMC 55 nm	314M	121 mm²2.6 M/mm²	32	600MHz	0.096	512MB GDDR3128-bit	28.8GB/s	50 W	–
GeForce 9500 GT Rev. 3	July 29, 2008	G96-300-C1	TSMC 55 nm	314M	121 mm²2.6 M/mm²	32	550MHz	0.090	512MB GDDR3128-bit	25.6GB/s	50 W	–
GeForce 9600 GSOEM	July 29, 2008	G94-300-A1	TSMC 65 nm	505M	240 mm²2.1 M/mm²	48	500MHz	0.120	768MB DDR2192-bit	48GB/s	–	–
GeForce 8400 SEOEM	August 1, 2008	G86S	TSMC 8 nm	210M	127 mm²1.7 M/mm²	16	459MHz	0.029	128MB / 256MB / 512MB DDR264-bit	6.4GB/s	50 W	–
GeForce 9400 GT	August 1 2008	G86S	TSMC 8 nm	210M	127 mm²1.7 M/mm²	16	459MHz	0.029	128MB / 256MB / 512MB DDR264-bit	9.6GB/s	50 W	–
GeForce 9400 GT (G96)	August 27, 2008	G96-200-C1	TSMC 55 nm	314M	121 mm²2.6 M/mm²	16	550MHz	0.045	512MB DDR2128-bit	12.8GB/s	50 W	–
GeForce 9400 GT PCI	August 27, 2008	G96-200-C1	TSMC 55 nm	314M	121 mm²2.6 M/mm²	16	550MHz	0.045	256MB DDR2128-bit	12.8GB/s	50 W	–
GeForce 9400 GT Rev. 2	August 27, 2008	G96-200-C1	TSMC 55 nm	314M	121 mm²2.6 M/mm²	16	550MHz	0.045	512MB / 1GB GDDR3128-bit	12.8GB/s	50 W	–
GeForce 9600 GSO 512	October 23, 2008	G94-300-A1 G94B	TSMC 65 nm	505M	240 mm²2.1 M/mm²	48	650MHz	0.156	512MB GDDR3256-bit	57.6GB/s	90 W	$189
GeForce GT 130 Mac Edition	December 12, 2008	G94B	TSMC 55 nm	505M	196 mm²2.6 M/mm²	48	600MHz	0.144	512MB GDDR3192-bit	38.0GB/s	75 W	–
GeForce 9600 GT Mac EditionOEM	December 22, 2008	G94-300-A1	TSMC 65 nm	505M	240 mm²2.1 M/mm²	64	600MHz	0.192	1GB DDR2256-bit	17.0GB/s	95 W	–
GeForce GT 120 Mac Edition	January 20, 2009	G96C	TSMC 55 nm	314M	121 mm²2.6 M/mm²	32	550MHz	0.090	512MB GDDR3128-bit	25.6GB/s	50 W	–
GeForce GTS 250	March 4, 2009	G92-426-B1	TSMC 55 nm	754M	260 mm²2.9 M/mm²	128	675MHz	0.415	1GB GDDR3256-bit	64.5GB/s	150 W	$199
GeForce G 100 OEMOEM	March 10, 2009	G98-309-U2	UMC 65 nm	210M	86 mm²2.4 M/mm²	8	540MHz	0.021	256MB DDR264-bit	6.4GB/s	35 W	–
GeForce GT 120 OEMOEM	March 10, 2009	G96C	TSMC 55 nm	314M	121 mm²2.6 M/mm²	32	738MHz	0.118	512MB DDR2128-bit	16.1GB/s	50 W	–
GeForce GT 130 OEMOEM	March 10, 2009	G94B	TSMC 55 nm	505M	196 mm²2.6 M/mm²	48	500MHz	0.120	512MB DDR2192-bit	24GB/s	75 W	–
GeForce GT 140 OEMOEM	March 10, 2009	G94B	TSMC 55 nm	505M	196 mm²2.6 M/mm²	64	650MHz	0.208	1GB GDDR3256-bit	57.6GB/s	105 W	–
GeForce GTS 150 OEMOEM	March 10, 2009	G92	TSMC 65 nm	754M	324 mm²2.3 M/mm²	128	738MHz	0.470	1GB GDDR3256-bit	64GB/s	141 W	–
GeForce GT 230 OEMOEM	April 27, 2009	G92-159-B1	TSMC 55 nm	754M	260 mm²2.9 M/mm²	96	500MHz	0.240	1.5GB DDR2192-bit	24GB/s	75 W	–
GeForce G210 OEMOEM	May 26, 2009	G96C	TSMC 55 nm	314M	121 mm²2.6 M/mm²	16	550MHz	0.043	512MB DDR264-bit	8.1GB/s	31 W	–
GeForce 9600 GTX	May 27, 2009	G92	TSMC 65 nm	754M	324 mm²2.3 M/mm²	96	580MHz	0.278	512MB GDDR3256-bit	44.8GB/s	140 W	–
GeForce GTS 240 OEMOEM	July 1, 2009	G92-240-B1	TSMC 55 nm	754M	260 mm²2.9 M/mm²	112	675MHz	0.363	1GB GDDR3256-bit	70.4GB/s	120 W	–
GeForce GT 230	October 12, 2009	G94-300-B1	TSMC 55 nm	505M	196 mm²2.6 M/mm²	48	650MHz	0.156	512MB / 1GB GDDR3256-bit	57.6GB/s	75 W	$44
GeForce GT 220 (G94)	January 26, 2010	G94	TSMC 65 nm	505M	240 mm²2.1 M/mm²	48	600MHz	0.144	1GB GDDR3128-bit	22.4GB/s	58 W	$49
GeForce GT 330 (G92)	February 2, 2010	G92	TSMC 65 nm	754M	324 mm²2.3 M/mm²	96	500MHz	0.240	256MB / 512MB / 768MB / 1GB / 2GB GDDR3256-bit	51.2GB/s	75 W	–
GeForce GTX 275 PhysX Edition	February 16, 2010	G92-428-B1 + G200-105-B3	TSMC 55 nm	754M	260 mm²2.9 M/mm²	128	633 / 738MHz	0.622	384MB / 896MB GDDR3448-bit	127GB/s	219 W	$349
GeForce 9500 GT Mac EditionOEM	May 19, 2010	G96C	TSMC 55 nm	314M	121 mm²2.6 M/mm²	32	550MHz	0.090	512MB GDDR3128-bit	25.6GB/s	50 W	–
Workstation 17
Quadro FX 4600	Mar 5, 2007	G80-850-A2 + NVIO-1-A3	TSMC 9 nm	681M	484 mm²1.4 M/mm²	96	500MHz	0.345	768MB GDDR3384-bit	67.2GB/s	134 W	–
Quadro FX 4600 SDI	Mar 5, 2007	G80-850-A2 + NVIO-1-A3	TSMC 9 nm	681M	484 mm²1.4 M/mm²	96	500MHz	0.345	768MB GDDR3384-bit	67.2GB/s	154 W	–
Quadro FX 5600	Mar 5, 2007	G80-875-A2 + NVIO-1-A3	TSMC 9 nm	681M	484 mm²1.4 M/mm²	128	600MHz	0.518	528MB GDDR3384-bit	76.8GB/s	171 W	–
Quadro FX 1700	Sep 12, 2007	G84-875-A2	TSMC 8 nm	289M	169 mm²1.7 M/mm²	32	460MHz	0.088	512MB DDR2128-bit	12.8GB/s	42 W	–
Quadro FX 370	Sep 12, 2007	G84-825-A2	TSMC 8 nm	289M	169 mm²1.7 M/mm²	16	360MHz	0.035	256MB DDR264-bit	6.4GB/s	35 W	–
Quadro FX 570	Sep 12, 2007	G84-850-A2	TSMC 8 nm	289M	169 mm²1.7 M/mm²	16	460MHz	0.044	256MB DDR2128-bit	12.8GB/s	38 W	–
NVS 290	Oct 4, 2007	G86-827-A2	TSMC 8 nm	210M	127 mm²1.7 M/mm²	16	460MHz	0.044	256MB DDR264-bit	6.4GB/s	21 W	–
Quadro FX 3700	Jan 8, 2008	G92-875-A2	TSMC 65 nm	754M	324 mm²2.3 M/mm²	112	500MHz	0.420	512MB GDDR3256-bit	51.2GB/s	78 W	–
Quadro VX 200	Jan 8, 2008	G92-851-A2	TSMC 65 nm	754M	324 mm²2.3 M/mm²	96	450MHz	0.324	512MB GDDR3256-bit	51.2GB/s	75 W	–
Quadro FX 4700X2	Apr 18, 2008	G92-880-A2	TSMC 65 nm	754M	324 mm²2.3 M/mm²	128	600MHz	0.576	GDDR3256-bit	51.2GB/s	226 W	–
Quadro FX 370 LP	Nov 6, 2008	G98	UMC 8 nm	210M	86 mm²2.4 M/mm²	8	540MHz	0.026	256MB DDR264-bit	8GB/s	25 W	–
NVS 450	Nov 11, 2008	G98	UMC 65 nm	210M	86 mm²2.4 M/mm²	8	550MHz	0.031	256MB GDDR364-bit	11.2GB/s	35 W	–
NVS 420	Jan 20, 2009	G98-850-U2	UMC 65 nm	210M	86 mm²2.4 M/mm²	8	550MHz	0.031	256MB GDDR364-bit	11.2GB/s	40 W	–
Quadro FX 1800	Mar 30, 2009	G94-876-B1	TSMC 65 nm	505M	240 mm²2.1 M/mm²	64	550MHz	0.264	768MB GDDR3192-bit	38.4GB/s	59 W	–
Quadro FX 380	Mar 30, 2009	G96-850-C1	TSMC 65 nm	314M	121 mm²2.6 M/mm²	16	450MHz	0.053	256MB GDDR3128-bit	22.4GB/s	34 W	–
Quadro FX 580	Apr 9, 2009	G96-875-C1	TSMC 65 nm	314M	121 mm²2.6 M/mm²	32	450MHz	0.108	512MB GDDR3128-bit	25.6GB/s	40 W	–
NVS 295	May 7, 2009	G98	UMC 65 nm	210M	86 mm²2.4 M/mm²	8	550MHz	0.031	256MB GDDR364-bit	11.2GB/s	23 W	–
Datacenter 3
C870 GPU Computing Module	May 2, 2007	G80	TSMC 9 nm	681M	484 mm²1.4 M/mm²	128	600MHz	0.346	1.5GB GDDR3384-bit	76.8GB/s	171 W	–
D870 Deskside Computer	May 2, 2007	G80	TSMC 9 nm	681M	484 mm²1.4 M/mm²	128	600MHz	0.691	1.5GB / 2GB GDDR3384-bit	76.8GB/s	520 W	–
S870 GPU Computing Server	May 2, 2007	G80	TSMC 9 nm	681M	484 mm²1.4 M/mm²	128	600MHz	1.38	1.5GB / 4GB GDDR3384-bit	76.8GB/s	–	–

Tesla 2.0

2007–2012 · TSMC 65 nm / TSMC 55 nm / TSMC 40 nm · 50 GPUs · up to 1.4B transistors

GT200 (GeForce GTX 280, June 2008) raised the shader count to 240 and added double-precision maths. At 576 mm² on 65 nm it was the largest GPU die made up to then. The 40 nm GT215 to GT218 parts later brought DirectX 10.1 to the low end of the GeForce 200 and 300 series.

GPU	Launch	Die	Node	Transistors	Die size	Shaders	Clock	FP32 TFLOPS	Memory	Bandwidth	TDP	Launch price
Desktop 37
GeForce 8400 GS PCI Rev.2	April 17, 2007	GT218S	TSMC 4 nm	260M	57 mm²4.6 M/mm²	16	520MHz	0.039	512MB DDR364-bit	8GB/s	40 W	$70
GeForce GTX 260	June 16, 2008	G200-100-A2	TSMC 65 nm	1.4B	576 mm²2.4 M/mm²	192	576MHz	0.477	896MB GDDR3448-bit	111.9GB/s	182 W	$449
GeForce GTX 280	June 16, 2008	G200-300-A2	TSMC 65 nm	1.4B	576 mm²2.4 M/mm²	240	602MHz	0.622	1GB GDDR3512-bit	141.7GB/s	236 W	$649
GeForce GTX 260 Rev. 2	July 23, 2008	G200-103-B2	TSMC 55 nm	1.4B	576 mm²2.4 M/mm²	192	576MHz	0.477	896MB GDDR3448-bit	111.9GB/s	182 W	$449
GeForce GTX 260 Core 216	September 16, 2008	G200-103-A2	TSMC 65 nm	1.4B	576 mm²2.4 M/mm²	216	576MHz	0.537	896MB GDDR3448-bit	111.9GB/s	182 W	$299
GeForce GTX 260 Core 216 Rev. 2	November 27, 2008	G200-103-B2	TSMC 55 nm	1.4B	576 mm²2.4 M/mm²	216	576MHz	0.537	896MB GDDR3448-bit	111.9GB/s	171 W	$299
GeForce GTX 275	January 15, 2009	G200-105-B3	TSMC 55 nm	1.4B	470 mm²3 M/mm²	240	633MHz	0.674	896MB GDDR3448-bit	127GB/s	219 W	$249
GeForce GTX 285	January 15, 2009	GT200-350-B3	TSMC 55 nm	1.4B	470 mm²3 M/mm²	240	648MHz	0.708	1GB GDDR3512-bit	159GB/s	204 W	$359
GeForce GTX 285 Mac EditionOEM	January 15, 2009	GT200-350-B3	TSMC 55 nm	1.4B	470 mm²3 M/mm²	240	648MHz	0.708	1GB GDDR3512-bit	159GB/s	204 W	–
GeForce GTX 2952 GPUs	January 18, 2009	G200-400-B3	TSMC 55 nm	1.4B	576 mm²2.4 M/mm²	240	576MHz	0.596	896MB GDDR3448-bit	111.9GB/s	289 W	$499
GeForce GTX 295 Single PCB2 GPUs	June 16, 2009	G200-401-B3	TSMC 55 nm	1.4B	576 mm²2.4 M/mm²	240	576MHz	0.596	896MB GDDR3448-bit	112.9GB/s	289 W	$499
GeForce GTX 285 X2	June 17, 2009	G200-350-B3	TSMC 55 nm	1.4B	470 mm²3 M/mm²	240	648MHz	0.708	2GB GDDR3512-bit	147.5GB/s	315 W	$1,399
GeForce G210 OEM Rev. 2OEM	August 24, 2009	GT218-200-B1	TSMC 4 nm	260M	57 mm²4.6 M/mm²	16	589MHz	0.045	128MB / 256MB / 512MB / 1GB DDR264-bit	6.4GB/s	31 W	–
GeForce 210 OEMOEM	September 4, 2009	GT216	TSMC 4 nm	486M	100 mm²4.9 M/mm²	16	475MHz	0.035	1GB DDR264-bit	6.4GB/s	31 W	–
GeForce 210	October 12, 2009	GT218-300-A2	TSMC 4 nm	260M	57 mm²4.6 M/mm²	16	520MHz	0.039	512MB DDR364-bit	6.4GB/s	31 W	$29
GeForce 210 PCI	October 12, 2009	GT216	TSMC 4 nm	486M	100 mm²4.9 M/mm²	16	475MHz	0.035	512MB DDR264-bit	6.4GB/s	31 W	$29
GeForce 210 Rev. 2	October 12, 2009	GT218-325-B1	TSMC 4 nm	260M	57 mm²4.6 M/mm²	16	520MHz	0.039	1GB DDR364-bit	6.4GB/s	31 W	$29
GeForce GT 220	October 12, 2009	GT216-055-A3	TSMC 4 nm	486M	100 mm²4.9 M/mm²	48	625MHz	0.131	512MB / 1GB DDR2128-bit	25.3GB/s	58 W	$49
GeForce GT 220 OEMOEM	October 12, 2009	GT215-450-A2	TSMC 4 nm	727M	144 mm²5 M/mm²	48	506MHz	0.097	512MB / 1GB GDDR3128-bit	22.4GB/s	58 W	–
GeForce GT 240	November 17, 2009	GT215-450-A2	TSMC 4 nm	727M	144 mm²5 M/mm²	96	550MHz	0.257	1GB GDDR5128-bit	54.4GB/s	69 W	$79
GeForce 205 OEMOEM	November 26, 2009	GT218	TSMC 4 nm	260M	57 mm²4.6 M/mm²	16	589MHz	0.045	512MB DDR264-bit	6.4GB/s	31 W	–
GeForce 310	November 27, 2009	GT218-300-A2	TSMC 4 nm	260M	57 mm²4.6 M/mm²	16	589MHz	0.045	512MB DDR264-bit	5.3GB/s	31 W	–
GeForce GTX 260 OEMOEM	December 8, 2009	G200-101-A2	TSMC 65 nm	1.4B	576 mm²2.4 M/mm²	192	518MHz	0.415	1.75GB GDDR3448-bit	112.9GB/s	150 W	–
GeForce GT 320	February 2, 2010	GT215-250-A2	TSMC 4 nm	727M	144 mm²5 M/mm²	72	540MHz	0.188	1GB GDDR3128-bit	25.3GB/s	43 W	–
GeForce GT 330	February 2, 2010	GT215-301-A3	TSMC 4 nm	727M	144 mm²5 M/mm²	96	540MHz	0.250	512MB / 1GB GDDR3128-bit	25.3GB/s	75 W	–
GeForce GT 340	February 2, 2010	GT215-301-A3	TSMC 4 nm	727M	144 mm²5 M/mm²	96	550MHz	0.257	1GB GDDR3128-bit	27.2GB/s	69 W	–
GeForce 315 (GT216)	March 31, 2010	GT216	TSMC 4 nm	486M	100 mm²4.9 M/mm²	48	475MHz	0.106	512MB DDR364-bit	12.6GB/s	33 W	–
GeForce 8400 GS Rev. 3	July 12, 2010	GT218S	TSMC 4 nm	260M	57 mm²4.6 M/mm²	16	520MHz	0.039	512MB DDR264-bit	6.4GB/s	25 W	$50
GeForce 405OEM	September 3, 2010	GT218-300-B1	TSMC 4 nm	260M	57 mm²4.6 M/mm²	16	589MHz	0.045	512MB / 1GB DDR364-bit	12.6GB/s	25 W	–
GeForce 405 (GT216)OEM	September 3, 2010	GT216	TSMC 4 nm	486M	100 mm²4.9 M/mm²	24	475MHz	0.053	512MB / 1GB DDR2128-bit	12.8GB/s	25 W	–
GeForce GT 415 OEMOEM	September 3, 2010	GT216-305-A3	TSMC 4 nm	486M	100 mm²4.9 M/mm²	48	625MHz	0.131	512MB / 1GB DDR3128-bit	10.7GB/s	32 W	–
GeForce 315	February 22, 2011	GT218-300-B1	TSMC 4 nm	260M	57 mm²4.6 M/mm²	16	589MHz	0.045	1GB DDR364-bit	9.6GB/s	33 W	–
GeForce 9300 GS Rev. 2OEM	March 12, 2011	GT218-670-B1	TSMC 4 nm	260M	57 mm²4.6 M/mm²	16	589MHz	0.045	256MB DDR264-bit	5.3GB/s	–	–
GeForce 315 (GT215)	April 28, 2011	GT215-450-A2	TSMC 4 nm	727M	144 mm²5 M/mm²	48	506MHz	0.097	128MB / 256MB / 512MB DDR264-bit	5.6GB/s	33 W	–
GeForce 9400 GT Rev. 3	June 13, 2012	GT218-670-B1	TSMC 4 nm	260M	57 mm²4.6 M/mm²	16	589MHz	0.045	128MB DDR264-bit	9.6GB/s	50 W	–
GeForce 505 OEMOEM	January 6, 2013	GT218-300-B1	TSMC 4 nm	260M	57 mm²4.6 M/mm²	16	589MHz	0.045	512MB / 1GB DDR364-bit	9.6GB/s	25 W	–
GeForce 505 OEM (GT216)OEM	February 17, 2013	GT216	TSMC 4 nm	486M	100 mm²4.9 M/mm²	24	615MHz	0.049	1GB DDR3128-bit	22.4GB/s	25 W	–
Workstation 7
Quadro CX	Nov 11, 2008	GT200GL + NVIO2	55 nm	–	–	192	602MHz	0.694	528MB GDDR3384-bit	76.8GB/s	150 W	–
Quadro FX 4800	Nov 11, 2008	G200-850-B3 + NVIO2-A2	TSMC 55 nm	1.4B	576 mm²2.4 M/mm²	192	602MHz	0.694	528MB GDDR3384-bit	76.8GB/s	150 W	–
Quadro FX 5800	Nov 11, 2008	G200-875-B2 + NVIO2-A2	TSMC 55 nm	1.4B	576 mm²2.4 M/mm²	240	610MHz	0.878	96MB GDDR3512-bit	102.4GB/s	189 W	–
Quadro FX 3800	Mar 30, 2009	G200-835-B3 + NVIO2-A2	TSMC 55 nm	1.4B	576 mm²2.4 M/mm²	192	600MHz	0.691	GDDR3256-bit	51.2GB/s	108 W	–
Quadro FX 380 LP	Dec 1, 2009	GT218GL	4 nm	–	–	16	589MHz	0.067	512MB GDDR364-bit	12.8GB/s	28 W	–
NVS 300	Jan 8, 2011	GT218	TSMC 4 nm	260M	57 mm²4.6 M/mm²	16	589MHz	0.067	512MB DDR364-bit	12.6GB/s	18 W	–
Quadro 400	Apr 5, 2011	GT216GL	4 nm	–	–	48	450MHz	0.108	512MB DDR364-bit	12.3GB/s	32 W	–
Datacenter 6
S1070 GPU Computing Server "400 configuration"	June 1, 2008	GT200	TSMC 65 nm	1.4B	576 mm²2.4 M/mm²	240	602MHz	2.49	4GB GDDR3512-bit	98.5GB/s	800 W	–
S1070 GPU Computing Server "500 configuration"	June 1, 2008	GT200	TSMC 65 nm	1.4B	576 mm²2.4 M/mm²	240	602MHz	2.77	4GB GDDR3512-bit	98.5GB/s	800 W	–
S1075 GPU Computing Server	June 1, 2008	GT200	TSMC 65 nm	1.4B	576 mm²2.4 M/mm²	240	602MHz	2.77	4GB GDDR3512-bit	98.5GB/s	–	–
Quadro Plex 2200 D2 Visual Computing System	July 25, 2008	GT200GL	–	–	–	240	648MHz	1.24	2GB / 4GB GDDR3512-bit	102.4GB/s	–	–
Quadro Plex 2200 S4 Visual Computing System	July 25, 2008	GT200GL	–	–	–	240	648MHz	2.49	4GB GDDR3512-bit	102.4GB/s	1200 W	–
C1060 GPU Computing Module	April 9, 2009	GT200	TSMC 65 nm	1.4B	576 mm²2.4 M/mm²	240	602MHz	0.622	4GB GDDR3512-bit	102.4GB/s	188 W	–

Fermi

2010–2013 · TSMC 40 nm · 74 GPUs · up to 3.1B transistors

Fermi (GeForce GTX 480, March 2010) rebuilt the GPU around compute: CUDA cores grouped into streaming multiprocessors, a real L1/L2 cache hierarchy, ECC memory support and hardware tessellation for DirectX 11, on TSMC 40 nm. The first GF100 ran hot and shipped with cores disabled; the GF110 respin in the GTX 580 enabled all 512.

GPU	Launch	Die	Node	Transistors	Die size	Shaders	Clock	FP32 TFLOPS	Memory	Bandwidth	TDP	Launch price
Desktop 57
GeForce GTX 470	March 26, 2010	GF100-275-A3	–	3.1B	529 mm²5.9 M/mm²	448	608MHz	1.09	1.25GB GDDR5320-bit	133.9GB/s	215 W	$349
GeForce GTX 480	March 26, 2010	GF100-375-A3	–	3.1B	529 mm²5.9 M/mm²	480	701MHz	1.35	1.5GB GDDR5384-bit	177.4GB/s	250 W	$499
GeForce GTX 465	May 31, 2010	GF100-030-A3	–	3.1B	529 mm²5.9 M/mm²	352	608MHz	0.855	1GB GDDR5256-bit	102.7GB/s	200 W	$279
GeForce GTX 460	July 12, 2010	GF104-300-KA-A1	–	2.0B	332 mm²5.9 M/mm²	336	675MHz	0.907	768MB / 1GB GDDR5192-bit	86.4GB/s	160 W	$199
GeForce GT 420 OEMOEM	September 3, 2010	GF108-200-A1	TSMC 4 nm	585M	116 mm²5 M/mm²	48	700MHz	0.134	1GB DDR3128-bit	28.8GB/s	50 W	–
GeForce GTS 450	September 13, 2010	GF106-250-KA-A1	–	1.2B	238 mm²4.9 M/mm²	192	783MHz	0.601	1GB GDDR5128-bit	57.7GB/s	106 W	$129
GeForce GTS 450 OEMOEM	September 13, 2010	GF106	–	1.2B	238 mm²4.9 M/mm²	144	783MHz	0.451	512MB GDDR5128-bit	57.7GB/s	106 W	–
GeForce GT 430	October 11, 2010	GF108-300-A1	TSMC 4 nm	585M	116 mm²5 M/mm²	96	700MHz	0.269	512MB / 1GB / 2GB / 4GB DDR364-bit	9.6GB/s	49 W	$79
GeForce GT 430 OEMOEM	October 11, 2010	GF108-400-A1	TSMC 4 nm	585M	116 mm²5 M/mm²	96	700MHz	0.269	2GB DDR3128-bit	25.6GB/s	49 W	–
GeForce GT 440 OEMOEM	October 11, 2010	GF106	–	1.2B	238 mm²4.9 M/mm²	144	810MHz	0.467	1.5GB DDR3192-bit	43.2GB/s	65 W	–
GeForce GTS 450 OEM 1.5GBOEM	October 11, 2010	GF106-250-KB-A1	–	1.2B	238 mm²4.9 M/mm²	144	790MHz	0.455	1.5GB GDDR5192-bit	96GB/s	106 W	–
GeForce GTX 460 OEMOEM	October 11, 2010	GF104-325-A1	–	2.0B	332 mm²5.9 M/mm²	336	650MHz	0.874	1GB GDDR5256-bit	108.8GB/s	150 W	–
GeForce GTX 580	November 9, 2010	GF110-375-A1	–	3.0B	520 mm²5.8 M/mm²	512	772MHz	1.58	1.5GB GDDR5384-bit	192.4GB/s	244 W	$499
GeForce GTX 460 SE	November 15, 2010	GF104-225-A1	–	2.0B	332 mm²5.9 M/mm²	288	650MHz	0.749	1GB GDDR5256-bit	108.8GB/s	150 W	$160
GeForce GTX 460 SE V2	November 15, 2010	GF114-400-A1	TSMC 4 nm	2.0B	332 mm²5.9 M/mm²	288	650MHz	0.749	768MB GDDR5192-bit	81.6GB/s	150 W	$160
GeForce GTX 570	December 7, 2010	GF110-275-A1	–	3.0B	520 mm²5.8 M/mm²	480	732MHz	1.41	1.25GB GDDR5320-bit	152GB/s	219 W	$349
GeForce GTX 570 Rev. 2	December 7, 2010	GF110-275-A1	–	3.0B	520 mm²5.8 M/mm²	480	732MHz	1.41	1.25GB GDDR5320-bit	152GB/s	219 W	$349
GeForce GTX 560 Ti	January 25, 2011	GF114-400-A1	TSMC 4 nm	2.0B	332 mm²5.9 M/mm²	384	823MHz	1.24	1GB GDDR5256-bit	128.3GB/s	170 W	$249
GeForce GTX 560 Ti X22 GPUs	January 25, 2011	GF114-400-A1	TSMC 4 nm	2.0B	332 mm²5.9 M/mm²	384	850MHz	1.31	1GB GDDR5256-bit	128.3GB/s	170 W	$519
GeForce GT 440	February 1, 2011	GF108-400-A1	TSMC 4 nm	585M	116 mm²5 M/mm²	96	810MHz	0.311	256MB / 512MB / 1GB / 2GB DDR3128-bit	28.8GB/s	65 W	$79
GeForce GT 440 Mac EditionOEM	February 9, 2011	GF108-400-A1	TSMC 4 nm	585M	116 mm²5 M/mm²	96	810MHz	0.311	1GB DDR3128-bit	21.3GB/s	65 W	–
GeForce GTX 560 Ti OEMOEM	March 8, 2011	GF114-400-A1	TSMC 4 nm	2.0B	332 mm²5.9 M/mm²	384	823MHz	1.24	1GB GDDR5256-bit	128.3GB/s	170 W	–
GeForce GTX 460 X22 GPUs	March 11, 2011	GF104-300-KB-A1	–	2.0B	332 mm²5.9 M/mm²	336	701MHz	0.942	1GB GDDR5256-bit	115.2GB/s	160 W	$350
GeForce GTS 450 Rev. 2	March 15, 2011	GF116-200-KA-A1	TSMC 4 nm	1.2B	238 mm²4.9 M/mm²	192	783MHz	0.601	1GB GDDR5128-bit	57.7GB/s	106 W	$129
GeForce GTX 550 Ti	March 15, 2011	GF116-400-A1	TSMC 4 nm	1.2B	238 mm²4.9 M/mm²	192	900MHz	0.691	1GB GDDR5192-bit	98.5GB/s	116 W	$149
GeForce GTX 5902 GPUs	March 24, 2011	GF110-351-A1	TSMC 4 nm	3.0B	520 mm²5.8 M/mm²	512	608MHz	1.24	1.5GB GDDR5384-bit	164GB/s	365 W	$699
GeForce GTX 560 Ti OEM (GF110)OEM	March 30, 2011	GF110	–	3.0B	520 mm²5.8 M/mm²	352	732MHz	1.03	1.25GB GDDR5256-bit	128.8GB/s	210 W	–
GeForce GT 520	April 13, 2011	GF119-300-A1	TSMC 4 nm	292M	79 mm²3.7 M/mm²	48	810MHz	0.156	1GB / 2GB DDR364-bit	14.4GB/s	29 W	$59
GeForce GT 530 OEMOEM	May 14, 2011	GF108-200-A1	TSMC 4 nm	585M	116 mm²5 M/mm²	96	700MHz	0.269	1GB / 2GB DDR3128-bit	28.8GB/s	50 W	–
GeForce GT 545	May 14, 2011	GF116	TSMC 4 nm	1.2B	238 mm²4.9 M/mm²	144	720MHz	0.415	1.5GB / 3GB DDR3192-bit	38.4GB/s	70 W	$149
GeForce GT 545 OEMOEM	May 14, 2011	GF116-110-KA-A1	TSMC 4 nm	1.2B	238 mm²4.9 M/mm²	144	871MHz	0.501	1GB GDDR5128-bit	64GB/s	105 W	–
GeForce GTX 555 OEMOEM	May 14, 2011	GF114-200-KB-A1	TSMC 4 nm	2.0B	332 mm²5.9 M/mm²	288	736MHz	0.848	1GB GDDR5192-bit	91.9GB/s	150 W	–
GeForce GTX 560	May 17, 2011	GF114-325-A1	TSMC 4 nm	2.0B	332 mm²5.9 M/mm²	336	810MHz	1.09	1GB GDDR5256-bit	128GB/s	150 W	$199
GeForce GTX 580 Rev. 2	June 6, 2011	GF110-380-A1	–	3.0B	520 mm²5.8 M/mm²	512	772MHz	1.58	1.5GB GDDR5384-bit	192.4GB/s	244 W	$499
GeForce GTX 460 V2	September 24, 2011	GF114-400-A1	TSMC 4 nm	2.0B	332 mm²5.9 M/mm²	336	779MHz	1.05	1GB GDDR5192-bit	96.2GB/s	160 W	$199
GeForce GTX 460 V2 E.S.	September 24, 2011	GF114-400-A1	TSMC 4 nm	2.0B	332 mm²5.9 M/mm²	336	779MHz	1.05	1.25GB GDDR5256-bit	128.3GB/s	160 W	–
GeForce 510 OEMOEM	September 29, 2011	GF119S	TSMC 4 nm	292M	79 mm²3.7 M/mm²	48	523MHz	0.100	1GB / 2GB DDR364-bit	14.4GB/s	25 W	–
GeForce GTX 560 OEMOEM	November 29, 2011	GF110-040-A1	–	3.0B	520 mm²5.8 M/mm²	384	552MHz	0.848	1.25GB GDDR5320-bit	128.3GB/s	150 W	–
GeForce GTX 560 Ti 448	November 29, 2011	GF110-270-A1	–	3.0B	520 mm²5.8 M/mm²	448	732MHz	1.31	1.25GB GDDR5320-bit	152GB/s	210 W	$289
GeForce GT 520 OEM (GF108)OEM	February 1, 2012	GF108-200-A1	TSMC 4 nm	585M	116 mm²5 M/mm²	48	700MHz	0.134	1GB / 2GB DDR364-bit	10.7GB/s	29 W	–
GeForce GTX 560 SE	February 20, 2012	GF114-200-KB-A1	TSMC 4 nm	2.0B	332 mm²5.9 M/mm²	288	736MHz	0.848	1GB GDDR5192-bit	91.9GB/s	150 W	$149
GeForce 605 OEMOEM	April 2, 2012	GF119-200-A1	TSMC 4 nm	292M	79 mm²3.7 M/mm²	48	523MHz	0.100	512MB / 1GB DDR364-bit	14.4GB/s	25 W	–
GeForce GT 610	April 2, 2012	GF119-300-A1	TSMC 4 nm	292M	79 mm²3.7 M/mm²	48	810MHz	0.156	512MB / 1GB DDR364-bit	10.7GB/s	29 W	$49
GeForce GT 620 OEMOEM	April 2, 2012	GF119S	TSMC 4 nm	292M	79 mm²3.7 M/mm²	48	810MHz	0.156	512MB / 1GB DDR364-bit	14.4GB/s	30 W	–
GeForce GT 640 OEM RebrandOEM	April 24, 2012	GF116-150-A1	TSMC 4 nm	1.2B	238 mm²4.9 M/mm²	144	720MHz	0.415	1.5GB DDR3192-bit	38.4GB/s	75 W	–
GeForce GT 645 OEMOEM	April 24, 2012	GF114	TSMC 4 nm	2.0B	332 mm²5.9 M/mm²	288	776MHz	0.894	1GB / 1.5GB GDDR5192-bit	91.9GB/s	102 W	–
GeForce 615 OEMOEM	May 15, 2012	GF108	TSMC 4 nm	585M	116 mm²5 M/mm²	48	660MHz	0.127	1GB DDR3128-bit	28.8GB/s	49 W	–
GeForce GT 620	May 15, 2012	GF108-100-KB-A1	TSMC 4 nm	585M	116 mm²5 M/mm²	96	700MHz	0.269	1GB DDR364-bit	14.4GB/s	49 W	$59
GeForce GT 630	May 15, 2012	GF108-400-A1	TSMC 4 nm	585M	116 mm²5 M/mm²	96	810MHz	0.311	1GB DDR3128-bit	28.8GB/s	65 W	$79
GeForce GTS 450 Rev. 3	July 11, 2012	GF116-200-KA-A1	TSMC 4 nm	1.2B	238 mm²4.9 M/mm²	144	783MHz	0.451	1GB DDR3128-bit	22.4GB/s	106 W	$129
GeForce GT 520 OEMOEM	August 20, 2012	GF119S	TSMC 4 nm	292M	79 mm²3.7 M/mm²	48	589MHz	0.135	1GB DDR364-bit	8GB/s	29 W	–
GeForce GT 610 OEMOEM	September 9, 2012	GF119S	TSMC 4 nm	292M	79 mm²3.7 M/mm²	48	550MHz	0.115	1GB DDR364-bit	8GB/s	29 W	–
GeForce GT 625 OEMOEM	February 18, 2013	GF119-400-A1	TSMC 4 nm	292M	79 mm²3.7 M/mm²	48	874MHz	0.168	1GB DDR364-bit	13.2GB/s	29 W	–
GeForce GT 705 OEMOEM	March 27, 2014	GF119-400-A1	TSMC 4 nm	292M	79 mm²3.7 M/mm²	48	874MHz	0.168	1GB DDR364-bit	13.2GB/s	29 W	–
GeForce GT 730 (GF108)	June 18, 2014	GF108-400-A1	TSMC 4 nm	585M	116 mm²5 M/mm²	96	700MHz	0.269	1GB / 2GB DDR3128-bit	28.8GB/s	49 W	$69
GeForce GT 710 (GF119)	January 26, 2016	GF119-300-A1	TSMC 4 nm	292M	79 mm²3.7 M/mm²	48	810MHz	0.155	1GB DDR364-bit	14.4GB/s	29 W	$35
GeForce GT 610 OEM (GF108)OEM	April 2, 2016	GF108	TSMC 4 nm	585M	116 mm²5 M/mm²	48	701MHz	0.135	2GB DDR364-bit	8GB/s	29 W	–
Workstation 9
Quadro 4000	Nov 2, 2010	GF100	TSMC 4 nm	3.1B	529 mm²5.9 M/mm²	256	475MHz	0.486	2GB GDDR5256-bit	89.6GB/s	142 W	–
Quadro 6000	Dec 10, 2010	GF100	TSMC 4 nm	3.1B	529 mm²5.9 M/mm²	448	574MHz	1.03	6GB GDDR5384-bit	144GB/s	204 W	–
Quadro 600	Dec 13, 2010	GF108GL	4 nm	–	–	96	640MHz	0.246	1GB DDR3128-bit	25.6GB/s	40 W	–
Quadro 2000	Dec 24, 2010	GF106GL (GF106-875)	4 nm	–	–	192	625MHz	0.480	1GB GDDR5128-bit	41.6GB/s	62 W	–
Quadro 5000	Feb 23, 2011	GF100	TSMC 4 nm	3.1B	529 mm²5.9 M/mm²	352	513MHz	0.722	2.5GB GDDR5320-bit	120GB/s	152 W	–
Quadro Plex 7000	July 25, 2011	GF100	TSMC 4 nm	3.1B	529 mm²5.9 M/mm²	512	574MHz	1.18	2GB / 6GB GDDR5384-bit	144GB/s	600 W	–
Quadro 7000	May 2, 2012	GF110	TSMC 4 nm	3.0B	520 mm²5.8 M/mm²	512	651MHz	1.33	6GB GDDR5384-bit	177GB/s	204 W	–
NVS 310	Jun 26, 2012	GF119	TSMC 4 nm	292M	79 mm²3.7 M/mm²	48	523MHz	0.100	512MB DDR364-bit	14GB/s	20 W	–
NVS 315	Mar 10, 2013	GF119	TSMC 4 nm	292M	79 mm²3.7 M/mm²	48	523MHz	–	DDR364-bit	14GB/s	20 W	–
Datacenter 8
C2050 GPU Computing Module	July 25, 2011	GF100	TSMC 4 nm	3.1B	529 mm²5.9 M/mm²	448	575MHz	1.03	3GB GDDR5384-bit	144GB/s	247 W	–
C2070 GPU Computing Module	July 25, 2011	GF100	TSMC 4 nm	3.1B	529 mm²5.9 M/mm²	448	575MHz	1.03	6GB GDDR5384-bit	144GB/s	247 W	–
C2075 GPU Computing Module	July 25, 2011	GF100	TSMC 4 nm	3.1B	529 mm²5.9 M/mm²	448	575MHz	1.03	6GB GDDR5384-bit	144GB/s	225 W	–
M2050 GPU Computing Module	July 25, 2011	GF100	TSMC 4 nm	3.1B	529 mm²5.9 M/mm²	448	575MHz	1.03	3GB GDDR5384-bit	148.4GB/s	225 W	–
M2070/M2070Q GPU Computing Module	July 25, 2011	GF100	TSMC 4 nm	3.1B	529 mm²5.9 M/mm²	448	575MHz	1.03	6GB GDDR5384-bit	150.3GB/s	225 W	–
M2090 GPU Computing Module	July 25, 2011	GF110	TSMC 4 nm	3.0B	520 mm²5.8 M/mm²	512	650MHz	1.33	6GB GDDR5384-bit	177.6GB/s	225 W	–
S2050 GPU Computing Server	July 25, 2011	GF100	TSMC 4 nm	3.1B	529 mm²5.9 M/mm²	448	575MHz	4.12	3GB / 4GB GDDR5384-bit	148.4GB/s	900 W	–
S2070 GPU Computing Server	July 25, 2011	GF100	TSMC 4 nm	3.1B	529 mm²5.9 M/mm²	448	575MHz	4.12	4GB / 6GB GDDR5384-bit	148.4GB/s	900 W	–

Kepler

2012–2014 · TSMC 28 nm · 59 GPUs · 1 benchmarked · up to 7.1B transistors

Kepler (GeForce GTX 680, March 2012) dropped Fermi's double-speed shader clock for far more, slower cores, roughly doubling performance per watt on TSMC 28 nm. It introduced GPU Boost and the NVENC video encoder. The big GK110 arrived in the Tesla K20 and the first GTX Titan.

GPU	Launch	Die	Node	Transistors	Die size	Shaders	Clock	FP32 TFLOPS	Memory	Bandwidth	TDP	Launch price
Desktop 43
GeForce GTX 680	March 22, 2012	GK104-400-A2	TSMC 28 nm	3.5B	294 mm²12 M/mm²	1,536	1058MHz	3.25	2GB GDDR5256-bit	192.3GB/s	195 W	$499
GeForce GT 630 OEMOEM	April 24, 2012	GK107-301-A2	TSMC 28 nm	1.3B	118 mm²10.8 M/mm²	192	875MHz	0.336	1GB / 2GB DDR3128-bit	28.5GB/s	50 W	–
GeForce GT 640 OEMOEM	April 24, 2012	GK107-320-A2	TSMC 28 nm	1.3B	118 mm²10.8 M/mm²	384	797MHz	0.612	1GB / 2GB / 4GB DDR3128-bit	28.5GB/s	50 W	–
GeForce GTX 6902 GPUs	May 3, 2012	GK104-355-A2	TSMC 28 nm	3.5B	294 mm²12 M/mm²	1,536	1019MHz	3.13	2GB GDDR5256-bit	192.3GB/s	300 W	$999
GeForce GTX 670	May 10, 2012	GK104-325-A2	TSMC 28 nm	3.5B	294 mm²12 M/mm²	1,344	980MHz	2.63	2GB GDDR5256-bit	192.3GB/s	170 W	$399
GeForce GT 640	June 5, 2012	GK107-300-A2	TSMC 28 nm	1.3B	118 mm²10.8 M/mm²	384	902MHz	0.693	2GB DDR3128-bit	28.5GB/s	65 W	$99
GeForce GTX 660 Ti	August 16, 2012	GK104-300-KD-A2	TSMC 28 nm	3.5B	294 mm²12 M/mm²	1,344	980MHz	2.63	2GB GDDR5192-bit	144.2GB/s	150 W	$299
GeForce GTX 660 OEMOEM	August 22, 2012	GK104-200-KD-A2	TSMC 28 nm	3.5B	294 mm²12 M/mm²	1,152	888MHz	2.05	1.5GB / 2GB / 3GB GDDR5192-bit	134.4GB/s	130 W	–
GeForce GTX 650	September 6, 2012	GK107-450-A2	TSMC 28 nm	1.3B	118 mm²10.8 M/mm²	384	1058MHz	0.813	1GB GDDR5128-bit	80GB/s	65 W	$109
GeForce GTX 660	September 6, 2012	GK106-400-A1	TSMC 28 nm	2.5B	221 mm²11.5 M/mm²	960	1032MHz	1.59	2GB GDDR5192-bit	144.2GB/s	140 W	$229
GeForce GTX 650 Ti	October 9, 2012	GK106-220-A1	TSMC 28 nm	2.5B	221 mm²11.5 M/mm²	768	928MHz	1.43	1GB GDDR5128-bit	86.4GB/s	110 W	$149
GeForce GTX TITAN	February 21, 2013	GK110-400-A1	TSMC 28 nm	7.1B	561 mm²12.6 M/mm²	2,688	876MHz	4.71	6GB GDDR5384-bit	288.4GB/s	250 W	$999
GeForce GTX 650 Ti Boost	March 26, 2013	GK106-240-A1	TSMC 28 nm	2.5B	221 mm²11.5 M/mm²	768	1032MHz	1.59	2GB GDDR5192-bit	144.2GB/s	134 W	$169
GeForce GTX 650 Ti OEMOEM	March 31, 2013	GK106	TSMC 28 nm	2.5B	221 mm²11.5 M/mm²	768	928MHz	1.43	2GB GDDR5128-bit	86.4GB/s	110 W	–
GeForce GTX 680 Mac EditionOEM	April 3, 2013	GK104-400-A2	TSMC 28 nm	3.5B	294 mm²12 M/mm²	1,536	1058MHz	3.25	2GB GDDR5256-bit	192.3GB/s	195 W	–
GeForce GTX 645 OEMOEM	April 22, 2013	GK106-200-A1	TSMC 28 nm	2.5B	221 mm²11.5 M/mm²	576	824MHz	0.949	1GB GDDR5128-bit	64GB/s	65 W	–
GeForce GTX 780	May 23, 2013	GK110-300-A1	TSMC 28 nm	7.1B	561 mm²12.6 M/mm²	2,304	902MHz	4.16	3GB GDDR5384-bit	288.4GB/s	250 W	$649
GeForce GT 630 Rev. 2	May 29, 2013	GK208-301-A1	TSMC 28 nm	1.0B	87 mm²11.7 M/mm²	384	902MHz	0.693	2GB DDR364-bit	14.4GB/s	25 W	$79
GeForce GT 640 Rev. 2	May 29, 2013	GK208-400-A1	TSMC 28 nm	1.0B	87 mm²11.7 M/mm²	384	1046MHz	0.803	1GB GDDR564-bit	40.1GB/s	49 W	$89
GeForce GTX 770	May 30, 2013	GK104-425-A2	TSMC 28 nm	3.5B	294 mm²12 M/mm²	1,536	1085MHz	3.33	2GB GDDR5256-bit	224.4GB/s	230 W	$329
GeForce GTX 760	June 25, 2013	GK104-225-A2	TSMC 28 nm	3.5B	294 mm²12 M/mm²	1,152	1032MHz	2.38	2GB GDDR5256-bit	192.3GB/s	170 W	$219
GeForce GTX 760 OEMOEM	June 27, 2013	GK104-200-KD-A2	TSMC 28 nm	3.5B	294 mm²12 M/mm²	1,152	888MHz	2.05	1.5GB GDDR5192-bit	134.4GB/s	130 W	–
GeForce GTX 760 OEM (August 25, 201)OEM	August 25, 2013	GK104-200-KD-A2	TSMC 28 nm	3.5B	294 mm²12 M/mm²	1,152	888MHz	2.05	1.5GB GDDR5256-bit	179.2GB/s	130 W	–
GeForce GTX 780 Rev. 2	September 10, 2013	GK110-300-B1	TSMC 28 nm	7.1B	561 mm²12.6 M/mm²	2,304	902MHz	4.16	3GB / 6GB GDDR5384-bit	288.4GB/s	250 W	$649
GeForce GTX 760 Ti OEMOEM	September 27, 2013	GK104-325-A2	TSMC 28 nm	3.5B	294 mm²12 M/mm²	1,344	980MHz	2.63	2GB GDDR5256-bit	192.3GB/s	170 W	–
GeForce GT 635 OEMOEM	October 1, 2013	GK208	TSMC 28 nm	1.0B	87 mm²11.7 M/mm²	384	967MHz	0.743	1GB DDR364-bit	14.4GB/s	35 W	–
GeForce GTX 780 Ti	November 7, 2013	GK110-425-B1	TSMC 28 nm	7.1B	561 mm²12.6 M/mm²	2,880	876 / 928MHz	5.35	3GB GDDR5384-bit	336GB/s	250 W	$699
GeForce GTX 760 X22 GPUs	November 19 2013	GK104-225-A2	TSMC 28 nm	3.5B	294 mm²12 M/mm²	1,152	1072MHz	4.94	2GB GDDR5256-bit	192.3GB/s	250 W	$600
GeForce GTX 650 (GK106)	November 27, 2013	GK106	TSMC 28 nm	2.5B	221 mm²11.5 M/mm²	384	1058MHz	0.813	1GB GDDR5128-bit	80GB/s	65 W	$109
GeForce GTX 750 Ti OEMOEM	February 18, 2014	GK106	TSMC 28 nm	2.5B	221 mm²11.5 M/mm²	960	1098MHz	2.11	2GB GDDR5256-bit	192.3GB/s	60 W	–
GeForce GTX TITAN Black	February 18, 2014	GK110-430-B1	TSMC 28 nm	7.1B	561 mm²12.6 M/mm²	2,880	980MHz	5.65	6GB GDDR5384-bit	336GB/s	250 W	$999
GeForce GT 710	March 27, 2014	GK208-203-B1	TSMC 28 nm	1.0B	87 mm²11.7 M/mm²	192	954MHz	0.366	1GB / 2GB DDR364-bit	12.8GB/s	19 W	$35
GeForce GT 710 OEMOEM	March 27, 2014	GK208B	TSMC 28 nm	1.0B	87 mm²11.7 M/mm²	192	797MHz	0.306	1GB DDR364-bit	14.4GB/s	–	–
GeForce GTX TITAN Z2 GPUs	May 28, 2014	GK110-350-B1	TSMC 28 nm	7.1B	561 mm²12.6 M/mm²	2,880	876MHz	10.1	2GB / 6GB GDDR5384-bit	336GB/s	375 W	$2,999
GeForce GT 740	May 29, 2014	GK107-425-A2	TSMC 28 nm	1.3B	118 mm²10.8 M/mm²	384	993MHz	0.763	1GB GDDR5128-bit	80.2GB/s	64 W	$89
GeForce GT 730	June 18, 2014	GK208-302-B1	TSMC 28 nm	1.0B	87 mm²11.7 M/mm²	384	902MHz	0.693	1GB GDDR564-bit	40.1GB/s	38 W	$69
GeForce GTX 660 Rev. 2	September 13, 2014	GK104	TSMC 28 nm	3.5B	294 mm²12 M/mm²	960	1032MHz	1.59	2GB GDDR5192-bit	144.2GB/s	140 W	$229
GeForce GT 720	September 29, 2014	GK208B	TSMC 28 nm	1.0B	87 mm²11.7 M/mm²	192	797MHz	0.306	1GB DDR364-bit	12.8GB/s	19 W	$49
GeForce GT 730 OEMOEM	April 9, 2015	GK107	TSMC 28 nm	1.3B	118 mm²10.8 M/mm²	384	902MHz	0.693	1GB GDDR564-bit	40.1GB/s	64 W	–
GeForce GT 740 OEMOEM	April 14 2015	GK106S	TSMC 28 nm	2.5B	221 mm²11.5 M/mm²	384	1006MHz	0.773	1GB GDDR5128-bit	80.2GB/s	65 W	–
GeForce GTX 760 OEM 2GBOEM	November 6, 2016	GK104-200-KD-A2	TSMC 28 nm	3.5B	294 mm²12 M/mm²	1,344	1046MHz	2.81	2GB GDDR5256-bit	211.2GB/s	170 W	–
GeForce GT 720 OEMOEM	October 29, 2017	GK107	TSMC 28 nm	1.3B	118 mm²10.8 M/mm²	192	993MHz	0.381	1GB DDR3128-bit	28.5GB/s	50 W	–
GeForce GT 1030 (GK107)	September 30, 2018	GK107	TSMC 28 nm	1.3B	118 mm²10.8 M/mm²	384	1058MHz	0.813	2GB GDDR5128-bit	80GB/s	65 W	–
Workstation 11
Quadro 410	Aug 7, 2012	GK107	TSMC 28 nm	1.3B	118 mm²10.8 M/mm²	192	706MHz	0.271	512MB DDR364-bit	14.4GB/s	38 W	–
Quadro K5000	Aug 17, 2012	GK104	TSMC 28 nm	3.5B	294 mm²12 M/mm²	1,536	706MHz	2.17	4GB GDDR5256-bit	172.8GB/s	122 W	–
NVS 510	Oct 23, 2012	GK107	TSMC 28 nm	1.3B	118 mm²10.8 M/mm²	192	797MHz	0.306	48MB DDR3128-bit	28.5GB/s	35 W	–
Quadro K2000	Mar 1, 2013	GK107	TSMC 28 nm	1.3B	118 mm²10.8 M/mm²	384	954MHz	0.733	2GB GDDR5128-bit	64GB/s	51 W	–
Quadro K2000D	Mar 1, 2013	GK107	TSMC 28 nm	1.3B	118 mm²10.8 M/mm²	384	954MHz	0.733	2GB GDDR5128-bit	64GB/s	51 W	–
Quadro K4000	Mar 1, 2013	GK106	TSMC 28 nm	2.5B	221 mm²11.5 M/mm²	768	811MHz	1.25	3GB GDDR5192-bit	134.8GB/s	80 W	–
Quadro K600	Mar 1, 2013	GK107	TSMC 28 nm	1.3B	118 mm²10.8 M/mm²	192	876MHz	0.336	1GB DDR3128-bit	28.5GB/s	41 W	–
Quadro K6000	Jul 23, 2013	GK110	TSMC 28 nm	7.1B	561 mm²12.6 M/mm²	2,880	902MHz	5.20	12GB GDDR5384-bit	288GB/s	225 W	–
Quadro K420	Jul 22, 2014	GK107	TSMC 28 nm	1.3B	118 mm²10.8 M/mm²	192	780MHz	0.300	1GB / 2GB DDR3128-bit	29GB/s	41 W	–
Quadro K4200	Jul 22, 2014	GK104	TSMC 28 nm	3.5B	294 mm²12 M/mm²	1,344	780MHz	2.10	4GB GDDR5256-bit	172.8GB/s	105 W	–
Quadro K5200	Jul 22, 2014	GK110B	28 nm	–	–	2,304	650MHz	3.00	8GB GDDR5256-bit	192GB/s	150 W	–
Datacenter 5
K10	May 1, 2012	GK104	TSMC 28 nm	3.5B	294 mm²12 M/mm²	1,536	–	4.58	2GB / 4GB GDDR5256-bit	160GB/s	225 W	–
K20	November 12, 2012	GK110	TSMC 28 nm	7.1B	561 mm²12.6 M/mm²	2,496	–	3.52	5GB GDDR5320-bit	208GB/s	225 W	–
K20X	November 12, 2012	GK110	TSMC 28 nm	7.1B	561 mm²12.6 M/mm²	2,688	–	3.94	6GB GDDR5384-bit	250GB/s	235 W	–
K40	October 8, 2013	GK110B	–	–	–	2,880	–	5.04	12GB GDDR5384-bit	288GB/s	235 W	–
K80	November 17, 2014	GK210	TSMC 28 nm	7.1B	561 mm²12.7 M/mm²	2,496	–	8.74	2GB / 12GB GDDR5384-bit	240GB/s	300 W	–

Maxwell

2014–2016 · TSMC 28 nm · 27 GPUs · 3 benchmarked · up to 8.0B transistors

Maxwell debuted in the GTX 750 Ti (February 2014) and scaled up in the GTX 980 that September. With TSMC's 20 nm node unavailable for large GPUs, NVIDIA stayed on 28 nm and found the gains in design: smaller SM partitions, a much larger L2 cache and tiled rasterisation roughly doubled efficiency again.

GPU	Launch	Die	Node	Transistors	Die size	Shaders	Clock	FP32 TFLOPS	Memory	Bandwidth	TDP	Launch price
Desktop 14
GeForce GTX 745 OEMOEM	February 18, 2014	GM107-220-A2	TSMC 28 nm	1.9B	148 mm²12.6 M/mm²	384	1033MHz	0.793	4GB DDR3128-bit	28.8GB/s	55 W	–
GeForce GTX 750	February 18, 2014	GM107-300-A2	TSMC 28 nm	1.9B	148 mm²12.6 M/mm²	512	1085MHz	1.11	1GB GDDR5128-bit	80.2GB/s	55 W	$119
GeForce GTX 750 Ti	February 18, 2014	GM107-400-A2	TSMC 28 nm	1.9B	148 mm²12.6 M/mm²	640	1085MHz	1.39	2GB GDDR5128-bit	86.4GB/s	60 W	$149
GeForce GTX 980	September 18, 2014	GM204-400-A1	TSMC 28 nm	5.2B	398 mm²13.1 M/mm²	2,048	1126 / 1216MHz	4.98	4GB GDDR5256-bit	224GB/s	165 W	$549
GeForce GTX 970	September 19, 2014	GM204-200-A1	TSMC 28 nm	5.2B	398 mm²13.1 M/mm²	1,664	1050 / 1178MHz	3.92	4GB GDDR5256-bit	224GB/s	145 W	$329
GeForce GTX 960	January 22, 2015	GM206-300-A1	TSMC 28 nm	2.9B	228 mm²12.9 M/mm²	1,024	1127 / 1178MHz	2.41	2GB GDDR5128-bit	112.2GB/s	120 W	$199
GeForce GTX TITAN X	March 17, 2015	GM200-400-A1	–	8.0B	601 mm²13.3 M/mm²	3,072	1000 / 1089MHz	6.69	12GB GDDR5384-bit	336.6GB/s	250 W	$999
GeForce GTX 980 Ti	June 2, 2015	GM200-310-A1	TSMC 28 nm	8.0B	601 mm²13.3 M/mm²	2,816	1000 / 1076MHz	6.06	6GB GDDR5384-bit	336GB/s	250 W	$649
GeForce GTX 950	August 20, 2015	GM206-250-A1	TSMC 28 nm	2.9B	228 mm²12.9 M/mm²	768	1024 / 1188MHz	1.83	2GB GDDR5128-bit	105.8GB/s	90 W	$159
GeForce GTX 750 (GM206)	November 17, 2015	GM206	TSMC 28 nm	2.9B	228 mm²12.9 M/mm²	512	1239MHz	1.27	2GB GDDR5128-bit	80.2GB/s	60 W	$149
GeForce GTX 960 OEMOEM	November 26, 2015	GM206	TSMC 28 nm	2.9B	228 mm²12.9 M/mm²	1,024	1176 / 1201MHz	2.46	4GB GDDR5128-bit	112.2GB/s	–	–
GeForce GTX 960 OEM (GM204)OEM	November 26, 2015	GM204	–	5.2B	398 mm²13.1 M/mm²	1,280	924 / 1038MHz	2.66	3GB GDDR5192-bit	120.3GB/s	–	–
GeForce GTX 950 OEMOEM	2016	GM206S	–	2.9B	228 mm²12.9 M/mm²	1,024	937 / 1203MHz	2.46	2GB GDDR5128-bit	80.2GB/s	–	–
GeForce GTX 950 Low Power	March 1, 2016	GM206-251-A1	TSMC 28 nm	2.9B	228 mm²12.9 M/mm²	768	1026 / 1190MHz	1.83	2GB GDDR5128-bit	105.8GB/s	75 W	$159
Workstation 8
Quadro K2200	Jul 22, 2014	GM107-875-A2	TSMC 28 nm	1.9B	148 mm²12.6 M/mm²	640	1046MHz	1.34	4GB GDDR5128-bit	80.2GB/s	68 W	–
Quadro K620	Jul 22, 2014	GM107-850	TSMC 28 nm	1.9B	148 mm²12.6 M/mm²	384	1000MHz	0.768	2GB DDR3128-bit	28.8GB/s	45 W	–
Quadro K1200	Jan 28, 2015	GM107-860	TSMC 28 nm	1.9B	148 mm²12.6 M/mm²	512	954MHz	1.08	4GB GDDR5128-bit	80.2GB/s	45 W	–
Quadro M6000	Mar 21, 2015	GM200GL GM200-880-A1	28 nm	–	–	3,072	988MHz	6.07	12GB / 24GB GDDR5384-bit	317GB/s	250 W	$4,200
Quadro M4000	Jun 29, 2015	GM204-850-A1	TSMC 28 nm	5.2B	398 mm²13.1 M/mm²	1,664	773MHz	2.66	8GB GDDR5256-bit	192.4GB/s	120 W	$791
Quadro M5000	Jun 29, 2015	GM204-875-A1	TSMC 28 nm	5.2B	398 mm²13.1 M/mm²	2,048	861MHz	4.30	8GB GDDR5256-bit	211.6GB/s	150 W	$2,857
NVS 810	Nov 4, 2015	GM107	TSMC 28 nm	1.9B	148 mm²12.6 M/mm²	512	1033MHz	1.06	48MB DDR364-bit	14.4GB/s	68 W	–
Quadro M2000	Apr 8, 2016	GM206-875-A1	TSMC 28 nm	2.9B	228 mm²12.9 M/mm²	768	796MHz	1.81	4GB GDDR5128-bit	105.8GB/s	75 W	$438
Datacenter 5
M6	August 30, 2015	GM204-995-A1	TSMC 28 nm	5.2B	398 mm²13.1 M/mm²	1,536	–	3.23	8GB GDDR5256-bit	147.2GB/s	75 W	–
M60	August 30, 2015	GM204-895-A1	TSMC 28 nm	5.2B	398 mm²13.1 M/mm²	2,048	–	9.65	2GB / 8GB GDDR5256-bit	160GB/s	225 W	–
M4	November 10, 2015	GM206	TSMC 28 nm	2.9B	228 mm²12.9 M/mm²	1,024	–	2.20	4GB GDDR5128-bit	88GB/s	50 W	–
M40	November 10, 2015	GM200	TSMC 28 nm	8.0B	601 mm²13.3 M/mm²	3,072	–	6.84	12GB / 24GB GDDR5384-bit	288GB/s	250 W	–
M10	May 18, 2016	GM107	TSMC 28 nm	1.9B	148 mm²12.6 M/mm²	640	–	5.29	4GB / 8GB GDDR5128-bit	83GB/s	225 W	–

Pascal

2016–2019 · TSMC 16 nm / Samsung 14 nm · 38 GPUs · 8 benchmarked · up to 15B transistors

Pascal ended four years on 28 nm with TSMC 16 nm FinFET (Samsung 14 nm for the smallest GP107 and GP108). The GTX 1080 (May 2016) introduced GDDR5X; the datacenter GP100 (Tesla P100) brought HBM2 memory and the first NVLink interconnect.

GPU	Launch	Die	Node	Transistors	Die size	Shaders	Clock	FP32 TFLOPS	Memory	Bandwidth	TDP	Launch price
Desktop 22
GeForce GTX 1080	May 27, 2016	GP104-400-A1	TSMC 16 nm	7.2B	314 mm²22.9 M/mm²	2,560	1607 / 1733MHz	8.87	8GB GDDR5X256-bit	320GB/s	180 W	$599
GeForce GTX 1070	June 10, 2016	GP104-200-A1	TSMC 16 nm	7.2B	314 mm²22.9 M/mm²	1,920	1506 / 1683MHz	6.46	8GB GDDR5256-bit	256GB/s	150 W	$379
GeForce GTX 1060 6GB	July 19, 2016	GP106-400-A1	TSMC 16 nm	4.4B	200 mm²22 M/mm²	1,280	1506 / 1708MHz	4.38	6GB GDDR5192-bit	192GB/s	120 W	$249
TITAN X (Pascal)	August 2, 2016	GP102-400-A1	TSMC 16 nm	12B	471 mm²25.5 M/mm²	3,584	1417 / 1531MHz	11.0	12GB GDDR5X384-bit	480GB/s	250 W	$1,200
TITAN X Pascal	August 2, 2016	GP102-400-A1	TSMC 16 nm	12B	471 mm²25.1 M/mm²	3,584	1417 / 1531MHz	11.0	12GB GDDR5X384-bit	480.4GB/s	250 W	–
GeForce GTX 1060 3GB	August 18, 2016	GP106-300-A1	TSMC 16 nm	4.4B	200 mm²22 M/mm²	1,152	1506 / 1708MHz	3.94	3GB GDDR5192-bit	192.2GB/s	120 W	$199
GeForce GTX 1050	October 25, 2016	GP107-300-A1	Samsung 14 nm	3.3B	132 mm²25 M/mm²	640	1354 / 1455MHz	1.86	2GB GDDR5128-bit	112.1GB/s	75 W	$109
GeForce GTX 1050 Ti	October 25, 2016	GP107-400-A1	Samsung 14 nm	3.3B	132 mm²25 M/mm²	768	1290 / 1392MHz	2.14	4GB GDDR5128-bit	112GB/s	75 W	$139
GeForce GTX 1060 3GB (GP104)	December 25, 2016	GP104-140-KA-A1	TSMC 16 nm	7.2B	314 mm²22.9 M/mm²	1,152	1506 / 1708MHz	3.94	3GB GDDR5192-bit	192.2GB/s	120 W	$199
GeForce GTX 1080 Ti	March 5, 2017	GP102-350-K1-A1	TSMC 16 nm	12B	471 mm²25.1 M/mm²	3,584	1480 / 1582MHz	11.3	11GB GDDR5X352-bit	484GB/s	250 W	$699
TITAN Xp	April 6, 2017	GP102-450-A1	TSMC 16 nm	12B	471 mm²25.1 M/mm²	3,840	1405 / 1582MHz	12.2	12GB GDDR5X384-bit	547.7GB/s	250 W	$1,200
GeForce GTX 1060 6GB 9Gbps	April 20, 2017	GP106-410-A1	TSMC 16 nm	4.4B	200 mm²22 M/mm²	1,280	1506 / 1709MHz	4.38	6GB GDDR5192-bit	216.7GB/s	120 W	$299
GeForce GTX 1080 (April 20, 2017)	April 20, 2017	GP104-410-A1	TSMC 16 nm	7.2B	314 mm²22.9 M/mm²	2,560	1607 / 1733MHz	8.87	8GB GDDR5X256-bit	352.3GB/s	180 W	$499
GeForce GT 1030 (May 17, 2017)	May 17, 2017	GP108-300-A1	Samsung 14 nm	1.8B	74 mm²24.3 M/mm²	384	1228 / 1468MHz	1.13	2GB GDDR564-bit	48.1GB/s	30 W	$69
GeForce GTX 1070 Ti	November 2, 2017	GP104-300-A1	TSMC 16 nm	7.2B	314 mm²22.9 M/mm²	2,432	1607 / 1683MHz	8.19	8GB GDDR5256-bit	256GB/s	180 W	$449
GeForce GTX 1060 OEMOEM	December 26, 2017	GP106-350-K3-A1	TSMC 16 nm	4.4B	200 mm²22 M/mm²	1,280	1506 / 1709MHz	4.38	5GB GDDR5160-bit	160.2GB/s	120 W	–
GeForce GTX 1060 6GB (GP104)	March 8, 2018	GP104-150-KA-A1	TSMC 16 nm	7.2B	314 mm²22.9 M/mm²	1,280	1506 / 1708MHz	4.37	6GB GDDR5192-bit	192.2GB/s	120 W	$299
GeForce GT 1030	March 12, 2018	GP108-310-A1	Samsung 14 nm	1.8B	74 mm²24.3 M/mm²	384	1152 / 1379MHz	1.06	2GB DDR464-bit	16.8GB/s	20 W	$79
GeForce GTX 1050 3GB	May 21, 2018	GP107-301-K1-A1	Samsung 14 nm	3.3B	132 mm²25 M/mm²	768	1392 / 1518MHz	2.33	3GB GDDR596-bit	84.1GB/s	75 W	$109
GeForce GTX 1060 6GB GDDR5X	October 18, 2018	GP104-150-KA-A1	TSMC 16 nm	7.2B	314 mm²22.9 M/mm²	1,280	1506 / 1709MHz	4.38	6GB GDDR5X192-bit	192.2GB/s	120 W	$299
GeForce GTX 1070 GDDR5X	December 4, 2018	GP104-200-A1	TSMC 16 nm	7.2B	314 mm²22.9 M/mm²	1,920	1506 / 1683MHz	6.46	8GB GDDR5X256-bit	256.3GB/s	150 W	$379
GeForce GT 1010 OEMOEM	January 13, 2021	GP108-200-A1	Samsung 14 nm	1.8B	74 mm²24.3 M/mm²	256	1228 / 1468MHz	0.752	2GB GDDR564-bit	48.1GB/s	30 W	–
Workstation 10
Quadro GP100	Oct 1, 2016	GP100-876-A1	TSMC 16 nm	15B	610 mm²25.1 M/mm²	3,584	1304 / 1442MHz	10.3	16GB HBM24096-bit	720GB/s	235 W	–
Quadro P5000	Oct 1, 2016	GP104-875-A1	TSMC 16 nm	7.2B	314 mm²22.9 M/mm²	2,560	1607 / 1733MHz	8.87	16GB GDDR5X256-bit	288GB/s	180 W	$2,499
Quadro P6000	Oct 1, 2016	GP102-875-A1	TSMC 16 nm	12B	471 mm²25.5 M/mm²	3,840	1506 / 1645MHz	11.8	24GB GDDR5X384-bit	432GB/s	250 W	$5,999
Quadro P2000	Feb 6, 2017	GP106-875	TSMC 16 nm	4.4B	200 mm²22 M/mm²	1,024	1076 / 1480MHz	3.01	5GB GDDR5160-bit	140GB/s	75 W	$585
Quadro P4000	Feb 6, 2017	GP104-850-A1	TSMC 16 nm	7.2B	314 mm²22.9 M/mm²	1,792	1202 / 1480MHz	5.30	8GB GDDR5256-bit	243GB/s	105 W	$815
Quadro P1000	Feb 7, 2017	GP107-860	Samsung 14 nm	3.3B	132 mm²25 M/mm²	640	1266 / 1481MHz	1.89	4GB GDDR5128-bit	82GB/s	47 W	$375
Quadro P400	Feb 7, 2017	GP107-825	Samsung 14 nm	3.3B	132 mm²25 M/mm²	256	1228 / 1252MHz	0.641	2GB GDDR564-bit	32GB/s	30 W	$120
Quadro P600	Feb 7, 2017	GP107-850	Samsung 14 nm	3.3B	132 mm²25 M/mm²	384	1329 / 1557MHz	1.20	2GB GDDR5128-bit	64GB/s	40 W	$178
Quadro P620	Feb 1, 2018	GP107-855	Samsung 14 nm	3.3B	132 mm²25 M/mm²	512	1266 / 1354MHz	1.49	2GB GDDR5128-bit	64GB/s	40 W	–
Quadro P2200	Jun 10, 2019	GP106-880-K1-A1	TSMC 16 nm	4.4B	200 mm²22 M/mm²	1,280	1000 / 1493MHz	3.82	5GB GDDR5X160-bit	200GB/s	75 W	–
Datacenter 6
P100 SXM2	April 5, 2016	GP100-890-A1	TSMC 16 nm	15B	610 mm²25.1 M/mm²	3,584	–	10.6	16GB HBM24096-bit	732GB/s	300 W	–
P100 PCIe 12GB	June 20, 2016	GP100	TSMC 16 nm	15B	610 mm²25.1 M/mm²	3,584	–	9.34	12GB HBM23072-bit	549GB/s	250 W	–
P100 PCIe 16GB	June 20, 2016	GP100	TSMC 16 nm	15B	610 mm²25.1 M/mm²	3,584	–	9.34	16GB HBM24096-bit	732GB/s	250 W	–
P4	September 13, 2016	GP104	TSMC 16 nm	7.2B	314 mm²22.9 M/mm²	2,560	–	5.44	8GB GDDR5256-bit	192GB/s	50 W	–
P40	September 13, 2016	GP102	TSMC 16 nm	12B	471 mm²25.5 M/mm²	3,840	–	11.8	24GB GDDR5384-bit	345.6GB/s	250 W	–
P6	March 24, 2017	GP104-995-A1	TSMC 16 nm	7.2B	314 mm²22.9 M/mm²	2,048	–	6.17	16GB GDDR5256-bit	192.2GB/s	90 W	–

Volta

2017–2018 · TSMC 12 nm · 6 GPUs · 1 benchmarked · up to 21B transistors

Volta (Tesla V100, 2017) added tensor cores, units built for the mixed-precision matrix maths of neural-network training, and started NVIDIA's AI datacenter line. The 815 mm² GV100 on TSMC 12 nm reached consumers only as the Titan V.

GPU	Launch	Die	Node	Transistors	Die size	Shaders	Clock	FP32 TFLOPS	Memory	Bandwidth	TDP	Launch price
Desktop 2
Titan V	December 7, 2017	GV100-400-A1	TSMC 12 nm	21B	815 mm²25.9 M/mm²	5,120	1200 / 1455MHz	14.9	12GB HBM23072-bit	652.8GB/s	250 W	$2,999
Titan V 32GB	June 21, 2018	GV100-???-A1	TSMC 12 nm	21B	815 mm²25.9 M/mm²	5,120	1200 / 1455MHz	14.9	32GB HBM24096-bit	868.4GB/s	250 W	$5,999
Workstation 1
Quadro GV100	Mar 27, 2018	GV100-875-A1	TSMC 12 nm	21B	815 mm²25.9 M/mm²	5,120	1132 / 1627MHz	14.8	32GB HBM24096-bit	870GB/s	250 W	–
Datacenter 3
V100 SXM2	May 10, 2017	GV100-895-A1	TSMC 12 nm	21B	815 mm²25.9 M/mm²	5,120	–	14.9	16GB / 32GB HBM24096-bit	900GB/s	300 W	–
V100 PCIe	June 21, 2017	GV100	TSMC 12 nm	21B	815 mm²25.9 M/mm²	5,120	–	14.0	16GB / 32GB HBM24096-bit	900GB/s	250 W	–
V100	March 27, 2018	GV100	TSMC 12 nm	21B	815 mm²25.9 M/mm²	5,120	–	13.2	16GB HBM24096-bit	829.4GB/s	250 W	–

Turing

2018–2022 · TSMC 12 nm · 27 GPUs · 17 benchmarked · up to 19B transistors

Turing (GeForce RTX 2080, September 2018) added RT cores for real-time ray tracing and brought tensor cores to GeForce for DLSS upscaling, on TSMC 12 nm. The GTX 16 series used smaller Turing dies without either.

GPU	Launch	Die	Node	Transistors	Die size	Shaders	Clock	FP32 TFLOPS	Memory	Bandwidth	TDP	Launch price
Desktop 19
GeForce RTX 2080	September 20, 2018	TU104-400A-A1	TSMC 12 nm	14B	545 mm²25 M/mm²	2,944	1515 / 1710MHz	10.1	8GB GDDR6256-bit	448GB/s	215 W	$699
GeForce RTX 2080 Ti	September 20, 2018	TU102-300A-K1-A1	TSMC 12 nm	19B	754 mm²24.7 M/mm²	4,352	1350 / 1545MHz	13.4	11GB GDDR6352-bit	616GB/s	250 W	$999
GeForce RTX 2070	October 17, 2018	TU106-400A-A1	TSMC 12 nm	11B	445 mm²24.3 M/mm²	2,304	1410 / 1620MHz	7.47	8GB GDDR6256-bit	448GB/s	175 W	$499
Titan RTX	December 18, 2018	TU102-400-A1	TSMC 12 nm	19B	754 mm²24.7 M/mm²	4,608	1350 / 1770MHz	16.3	24GB GDDR6384-bit	672GB/s	280 W	$2,499
GeForce RTX 2060	January 7, 2019	TU106-200A-KA-A1	TSMC 12 nm	11B	445 mm²24.3 M/mm²	1,920	1365 / 1680MHz	6.45	6GB GDDR6192-bit	336GB/s	160 W	$349
GeForce GTX 1660 Ti	February 21, 2019	TU116-400-A1	TSMC 12 nm	6.6B	284 mm²23.2 M/mm²	1,536	1500 / 1770MHz	5.44	6GB GDDR6192-bit	288GB/s	120 W	$279
GeForce GTX 1660	March 14, 2019	TU116-300-A1	TSMC 12 nm	6.6B	284 mm²23.2 M/mm²	1,408	1530 / 1785MHz	5.03	6GB GDDR5192-bit	192GB/s	120 W	$219
GeForce GTX 1650	April 23, 2019	TU117-300-A1	TSMC 12 nm	4.7B	200 mm²23.5 M/mm²	896	1485 / 1665MHz	2.98	4GB GDDR5/GDDR6128-bit	128GB/s	75 W	$149
GeForce RTX 2060 SUPER	July 9, 2019	TU106-410-A1	TSMC 12 nm	11B	445 mm²24.3 M/mm²	2,176	1470 / 1650MHz	7.18	8GB GDDR6256-bit	448GB/s	175 W	$399
GeForce RTX 2070 SUPER	July 9, 2019	TU104-410-A1	TSMC 12 nm	14B	545 mm²25 M/mm²	2,560	1605 / 1770MHz	9.06	8GB GDDR6256-bit	448GB/s	215 W	$499
GeForce RTX 2080 SUPER	July 23, 2019	TU104-450-A1	TSMC 12 nm	14B	545 mm²25 M/mm²	3,072	1650 / 1815MHz	11.2	8GB GDDR6256-bit	496GB/s	250 W	$699
GeForce GTX 1660 Super	October 29, 2019	TU116-300-A1	TSMC 12 nm	6.6B	284 mm²23.2 M/mm²	1,408	1530 / 1785MHz	5.03	6GB GDDR6192-bit	336GB/s	125 W	$229
GeForce GTX 1650 Super	November 22, 2019	TU116-250-KA-A1	TSMC 12 nm	6.6B	284 mm²23.2 M/mm²	1,280	1530 / 1725MHz	4.42	4GB GDDR6128-bit	192GB/s	100 W	$159
GeForce RTX 2060 (TU104)	January 10, 2020	TU104-150-KC-A1	–	14B	545 mm²25 M/mm²	1,920	1365 / 1680MHz	6.45	6GB GDDR6192-bit	336GB/s	160 W	$299
GeForce GTX 1650 (April 1, 2020)	April 1, 2020	TU117-300-A1	–	4.7B	200 mm²23.5 M/mm²	896	1410 / 1590MHz	2.85	4GB GDDR6128-bit	192GB/s	75 W	$149
GeForce GTX 1650 (TU106)	June 18, 2020	TU106-125-KAB-A1	–	11B	445 mm²24.3 M/mm²	896	1410 / 1590MHz	2.85	4GB GDDR6128-bit	192GB/s	90 W	$149
GeForce GTX 1650 (TU116)	July 7, 2020	TU116-150-KA-A1	–	6.6B	284 mm²23.2 M/mm²	896	1410 / 1590MHz	2.85	4GB GDDR6128-bit	192GB/s	80 W	$149
GeForce RTX 2060 12GB	December 7, 2021	TU106-300-KA-A1	–	11B	445 mm²24.3 M/mm²	2,176	1470 / 1650MHz	7.18	12GB GDDR6192-bit	336GB/s	184 W	$299
GeForce GTX 1630	June 28, 2022	TU117-150-KA-A1	–	4.7B	200 mm²23.5 M/mm²	512	1740 / 1785MHz	1.83	4GB GDDR664-bit	96GB/s	75 W	$149
Workstation 7
Quadro RTX 5000	Aug 13, 2018	TU104-875-A1	TSMC 12 nm	14B	545 mm²25 M/mm²	3,072	1620 / 1770MHz	11.2	16GB GDDR6256-bit	448GB/s	230 W	$2,299
Quadro RTX 6000	Aug 13, 2018	TU102-875-A1	TSMC 12 nm	19B	754 mm²24.7 M/mm²	4,608	1440 / 1770MHz	16.3	24GB GDDR6384-bit	672GB/s	260 W	$6,299
Quadro RTX 8000	Aug 13, 2018	TU102-875-A1	TSMC 12 nm	19B	754 mm²24.7 M/mm²	4,608	1395 / 1770MHz	16.3	48GB GDDR6384-bit	672GB/s	260 W	$9,999
Quadro RTX 4000	Nov 13, 2018	TU104-850-A1	TSMC 12 nm	14B	545 mm²25 M/mm²	2,304	1005 / 1545MHz	7.12	8GB GDDR6256-bit	416GB/s	100 W	$899
T600	Apr 12, 2021	TU117-850-A1	TSMC 12 nm	4.7B	200 mm²23.5 M/mm²	640	735 / 1335MHz	1.70	4GB GDDR6128-bit	160GB/s	40 W	–
T1000	May 6, 2021	TU117	TSMC 12 nm	4.7B	200 mm²23.5 M/mm²	896	1065 / 1395MHz	2.50	4GB / 8GB GDDR6128-bit	160GB/s	50 W	–
T400	May 6, 2021	TU117	TSMC 12 nm	4.7B	200 mm²23.5 M/mm²	384	420 / 1425MHz	1.09	2GB / 4GB GDDR664-bit	80GB/s	30 W	–
Datacenter 1
T4	September 12, 2018	TU104-895-A1	TSMC 12 nm	14B	545 mm²25 M/mm²	2,560	585 / 1590MHz	8.10	16GB GDDR6256-bit	320GB/s	70 W	$2,299

Ampere

2020–2024 · Samsung 8 nm / TSMC 7 nm · 40 GPUs · 23 benchmarked · up to 54B transistors

Ampere came in two forms: GA100 for the datacenter (A100, May 2020) on TSMC 7 nm, and GA102 to GA107 on Samsung 8 nm for GeForce RTX 30 and the RTX A workstation cards, which doubled FP32 throughput per streaming multiprocessor.

GPU	Launch	Die	Node	Transistors	Die size	Shaders	Clock	FP32 TFLOPS	Memory	Bandwidth	TDP	Launch price
Desktop 21
GeForce RTX 3080	September 17, 2020	GA102-200-KD-A1	Samsung 8 nm	28B	628.4 mm²45 M/mm²	8,704	1440 / 1710MHz	29.8	10GB GDDR6X320-bit	760GB/s	320 W	$699
GeForce RTX 3090	September 24, 2020	GA102-300-A1	Samsung 8 nm	28B	628.4 mm²45 M/mm²	10,496	1395 / 1695MHz	35.6	24GB GDDR6X384-bit	936.2GB/s	350 W	$1,499
GeForce RTX 3070	October 29, 2020	GA104-300-A1	Samsung 8 nm	17B	392.5 mm²44.3 M/mm²	5,888	1500 / 1725MHz	20.3	8GB GDDR6256-bit	448GB/s	220 W	$499
GeForce RTX 3060 Ti	December 2, 2020	GA104-200-A1	Samsung 8 nm	17B	392.5 mm²44.3 M/mm²	4,864	1410 / 1665MHz	16.2	8GB GDDR6256-bit	448GB/s	200 W	$399
GeForce RTX 3060 12GB	February 25, 2021	GA106-300-A1	Samsung 8 nm	12B	276 mm²43.5 M/mm²	3,584	1320 / 1777MHz	12.7	12GB GDDR6192-bit	360GB/s	170 W	$329
GeForce RTX 3060 OEMOEM	February 25, 2021	GA106-400-A1	–	12B	276 mm²43.5 M/mm²	3,840	1627 / 1852MHz	14.2	6GB GDDR6192-bit	336GB/s	185 W	–
GeForce RTX 3080 Ti	June 3, 2021	GA102-225-A1	Samsung 8 nm	28B	628.4 mm²45 M/mm²	10,240	1365 / 1665MHz	34.1	12GB GDDR6X384-bit	912.4GB/s	350 W	$1,199
GeForce RTX 3070 Ti	June 10, 2021	GA104-400-A1	Samsung 8 nm	17B	392.5 mm²44.3 M/mm²	6,144	1575 / 1770MHz	21.8	8GB GDDR6X256-bit	608GB/s	290 W	$599
GeForce RTX 3060 12GB (GA104)	September 1, 2021	GA104-150-KC-A1	–	17B	392.5 mm²44.3 M/mm²	3,584	1320 / 1777MHz	12.7	12GB GDDR6192-bit	360GB/s	170 W	$329
GeForce RTX 3080 Ti 20GB	January 2022	GA102-250-KD-A1	–	28B	628.4 mm²45 M/mm²	10,240	1335 / 1665MHz	34.1	20GB GDDR6X320-bit	760.3GB/s	350 W	$1,199
GeForce RTX 3050 (GA106)	January 4, 2022	GA106-150-KA-A1	–	12B	276 mm²43.5 M/mm²	2,560	1552 / 1777MHz	9.10	8GB GDDR6128-bit	224GB/s	130 W	$249
GeForce RTX 3080 12GB	January 11, 2022	GA102-220-A1	Samsung 8 nm	28B	628.4 mm²45 M/mm²	8,960	1260 / 1710MHz	30.6	12GB GDDR6X384-bit	912GB/s	350 W	$799
GeForce RTX 3050 OEMOEM	January 27, 2022	GA106-150-KA-A1	–	12B	276 mm²43.5 M/mm²	2,304	1515 / 1755MHz	8.09	8GB GDDR6128-bit	224GB/s	130 W	–
GeForce RTX 3060 Ti (GA103)	February 23, 2022	GA103-200-A1	–	22B	496 mm²44.4 M/mm²	4,864	1410 / 1665MHz	16.2	8GB GDDR6256-bit	448GB/s	200 W	$399
GeForce RTX 3090 Ti	March 29, 2022	GA102-350-A1	Samsung 8 nm	28B	628.4 mm²45 M/mm²	10,752	1560 / 1860MHz	40.0	24GB GDDR6X384-bit	1,008GB/s	450 W	$1,999
GeForce RTX 3060 (GA104)	October 2022	GA104	–	17B	392.5 mm²44.3 M/mm²	3,584	1320 / 1777MHz	12.7	8GB GDDR6128-bit	240GB/s	195 W	$329
GeForce RTX 3060	October 12, 2022	GA106-302-A1	–	12B	276 mm²43.5 M/mm²	3,584	1320 / 1777MHz	12.7	8GB GDDR6128-bit	240GB/s	170 W	$329
GeForce RTX 3060 Ti GDDR6X	October 19, 2022	GA104-202-A1	–	17B	392.5 mm²44.3 M/mm²	4,864	1410 / 1665MHz	16.2	8GB GDDR6X256-bit	608.3GB/s	225 W	$399
GeForce RTX 3070 Ti (GA102)	October 21, 2022	GA102-150-A1	–	28B	628.4 mm²45 M/mm²	6,144	1575 / 1770MHz	21.8	8GB GDDR6X256-bit	608.3GB/s	290 W	$599
GeForce RTX 3050	December 16, 2022	GA107-350-A1	Samsung 8 nm	8.7B	200 mm²43.5 M/mm²	2,560	1530 / 1777MHz	9.10	8GB GDDR6128-bit	224GB/s	130 W	$249
GeForce RTX 3050 6GB	February 2, 2024	GA107-325-K2-A1	–	8.7B	200 mm²43.5 M/mm²	2,304	1042 / 1470MHz	6.77	6GB GDDR696-bit	168GB/s	70 W	$179
Workstation 8
RTX A6000	Oct 5, 2020	GA102-875-A1	Samsung 8 nm	28B	628.4 mm²45 M/mm²	10,752	1410 / 1860MHz	38.7	48GB GDDR6384-bit	768GB/s	300 W	$4,649
RTX A4000	Apr 12, 2021	GA104-875-A1	Samsung 8 nm	17B	392.5 mm²44.3 M/mm²	6,144	735 / 1560MHz	19.2	16GB GDDR6256-bit	448GB/s	140 W	$1,000
RTX A5000	Apr 12, 2021	GA102-850-A1	Samsung 8 nm	28B	628.4 mm²45 M/mm²	8,192	1395 / 1695MHz	27.8	24GB GDDR6384-bit	768GB/s	230 W	$2,250
RTX A2000	Aug 10, 2021	GA106-850-A1	Samsung 8 nm	12B	276 mm²43.5 M/mm²	3,328	562 / 1200MHz	7.99	6GB GDDR6192-bit	288GB/s	70 W	$449
RTX A4500	Nov 23, 2021	GA102-825-KD-A1	Samsung 8 nm	28B	628.4 mm²45 M/mm²	7,168	1200 / 1575MHz	23.7	20GB GDDR6320-bit	640GB/s	200 W	$1,199
RTX A5500	Mar 22, 2022	GA102-860-A1	Samsung 8 nm	28B	628.4 mm²45 M/mm²	10,240	1080 / 1665MHz	34.1	24GB GDDR6384-bit	768GB/s	230 W	$3,600
RTX A1000	Apr 16, 2024	GA107	Samsung 8 nm	8.7B	200 mm²43.5 M/mm²	2,304	727 / 1462MHz	6.74	8GB GDDR6128-bit	192GB/s	50 W	$365
RTX A400	Apr 16, 2024	GA107	Samsung 8 nm	8.7B	200 mm²43.5 M/mm²	768	727 / 1762MHz	2.71	4GB GDDR664-bit	96GB/s	50 W	$135
Datacenter 11
A100	May 14, 2020	GA100-883AA-A1	TSMC 7 nm	54B	826 mm²65.6 M/mm²	6,912	765 / 1410MHz	19.5	40GB HBM25120-bit	1,555GB/s	250 W	$10,000
A100 40GB SXM4	May 14, 2020	GA100-883AA-A1	TSMC 7 nm	54B	826 mm²65.6 M/mm²	6,912	1095 / 1410MHz	19.5	40GB HBM25120-bit	1,555GB/s	400 W	$12,000
A40	October 5, 2020	GA102	Samsung 8 nm	28B	628.4 mm²45 M/mm²	10,752	1305 / 1740MHz	37.4	48GB GDDR6384-bit	696GB/s	300 W	$5,000
A100 80GB SXM4	November 16, 2020	GA100-883AA-A1	TSMC 7 nm	54B	826 mm²65.6 M/mm²	6,912	1275 / 1410MHz	19.5	80GB HBM2e5120-bit	2,039GB/s	400 W	$17,000
A10	April 12, 2021	GA102-890-A1	Samsung 8 nm	28B	628.4 mm²45 M/mm²	9,216	–	31.2	24GB GDDR6384-bit	600GB/s	150 W	–
A10G	April 12, 2021	GA102	Samsung 8 nm	28B	628.4 mm²45 M/mm²	9,216	1320 / 1710MHz	31.5	24GB GDDR6384-bit	600GB/s	150 W	$2,800
A16	April 12, 2021	GA107	Samsung 8 nm	8.7B	200 mm²43.5 M/mm²	1,280	–	4.61	4GB / 16GB GDDR6128-bit	200GB/s	250 W	–
A30	April 12, 2021	GA100	TSMC 7 nm	54B	826 mm²65.6 M/mm²	3,584	–	10.3	24GB HBM23072-bit	933.1GB/s	165 W	–
A100 80GB PCIe	June 28, 2021	GA100-883AA-A1	TSMC 7 nm	54B	826 mm²65.6 M/mm²	6,912	1065 / 1410MHz	19.5	80GB HBM2e5120-bit	1,935GB/s	300 W	$15,000
A2	November 10, 2021	GA107	Samsung 8 nm	8.7B	200 mm²43.5 M/mm²	1,280	–	4.53	16GB GDDR6128-bit	200GB/s	40 W	–
A800 80GB	November 8, 2022	GA100-883AA-A1	TSMC 7 nm	54B	826 mm²65.6 M/mm²	6,912	1155 / 1410MHz	19.5	80GB HBM2e5120-bit	2,039GB/s	400 W	$15,000

Hopper

2022–2024 · TSMC 4 nm · 8 GPUs · 6 benchmarked · up to 80B transistors

Hopper is datacenter-only. The H100 (2022) on TSMC 4N added a Transformer Engine with FP8 maths, HBM3 memory and fourth-generation NVLink. The H200 kept the GH100 die and moved to 141 GB of HBM3e.

GPU	Launch	Die	Node	Transistors	Die size	Shaders	Clock	FP32 TFLOPS	Memory	Bandwidth	TDP	Launch price
H100 PCIe	March 22, 2022	GH100	TSMC 4 nm	80B	814 mm²98.3 M/mm²	14,592	1095 / 1755MHz	51.2	80GB HBM2e5120-bit	2,000GB/s	350 W	$25,000
H100 SXM	March 22, 2022	GH100	TSMC 4 nm	80B	814 mm²98.3 M/mm²	16,896	1590 / 1980MHz	66.9	80GB HBM35120-bit	3,350GB/s	700 W	$30,000
H100 NVL	March 21, 2023	GH100	TSMC 4 nm	80B	814 mm²98.3 M/mm²	16,896	1080 / 1785MHz	60.3	94GB HBM36016-bit	3,938GB/s	400 W	$29,000
H800	March 21, 2023	GH100	TSMC 4 nm	80B	814 mm²98.3 M/mm²	16,896	–	59.3	80GB HBM35120-bit	3,360GB/s	700 W	–
H800 80GB	March 21, 2023	GH100	TSMC 4 nm	80B	814 mm²98.3 M/mm²	16,896	1095 / 1980MHz	66.9	80GB HBM35120-bit	3,350GB/s	700 W	$35,000
GH200 Grace Hopper	May 28, 2023	GH100	TSMC 4 nm	80B	814 mm²98.3 M/mm²	16,896	1980MHz	66.9	141GB HBM3e6144-bit	4,900GB/s	700 W	$40,000
H200 PCIe	November 18, 2024	GH100	TSMC 4 nm	80B	814 mm²98.3 M/mm²	16,896	–	60.3	141GB HBM3E5120-bit	4,800GB/s	600 W	–
H200 SXM	November 18, 2024	GH100	TSMC 4 nm	80B	814 mm²98.3 M/mm²	16,896	1500 / 1980MHz	66.9	141GB HBM3e6144-bit	4,800GB/s	700 W	$31,000

Ada Lovelace

2022–2024 · TSMC 4 nm · 27 GPUs · 18 benchmarked · up to 76B transistors

Ada Lovelace (GeForce RTX 4090, October 2022) moved GeForce to TSMC 4N, with 76.3 billion transistors on the 608 mm² AD102. A much larger L2 cache offset narrower memory buses, and fourth-generation tensor cores enabled DLSS 3 frame generation.

GPU	Launch	Die	Node	Transistors	Die size	Shaders	Clock	FP32 TFLOPS	Memory	Bandwidth	TDP	Launch price
Desktop 17
GeForce RTX 4090	October 12, 2022	AD102-300-A1	TSMC 4 nm	76B	608.5 mm²125.4 M/mm²	16,384	2235 / 2520MHz	82.6	24GB GDDR6X384-bit	1,008GB/s	450 W	$1,599
GeForce RTX 4080	November 16, 2022	AD103-300-A1	TSMC 4 nm	46B	378.6 mm²121.2 M/mm²	9,728	2205 / 2505MHz	48.7	16GB GDDR6X256-bit	716.8GB/s	320 W	$1,199
RTX 4000 (Ada Generation)	2023	AD104-???	TSMC 4 nm	36B	294.5 mm²121.6 M/mm²	6,144	1500 / 2175MHz	26.7	20GB GDDR6 ECC160-bit	360GB/s	130 W	$1,250
GeForce RTX 4070 Ti	January 5, 2023	AD104-400-A1	TSMC 4 nm	36B	294.5 mm²121.6 M/mm²	7,680	2310 / 2610MHz	40.1	12GB GDDR6X192-bit	504GB/s	285 W	$799
GeForce RTX 4070	April 13, 2023	AD104-250-A1	TSMC 4 nm	36B	294.5 mm²121.6 M/mm²	5,888	1920 / 2475MHz	29.1	12GB GDDR6X192-bit	504GB/s	200 W	$599
GeForce RTX 4060 Ti	May 24, 2023	AD106-350-A1	–	23B	187.8 mm²121.9 M/mm²	4,352	2310 / 2535MHz	22.1	8GB GDDR6128-bit	288GB/s	160 W	$399
GeForce RTX 4060	June 29, 2023	AD107-400-A1	TSMC 4 nm	19B	158.7 mm²119.1 M/mm²	3,072	1830 / 2475MHz	15.1	8GB GDDR6128-bit	272GB/s	115 W	$299
GeForce RTX 4060 Ti 16GB	July 18, 2023	AD106-351-A1	TSMC 4 nm	23B	187.8 mm²121.9 M/mm²	4,352	2310 / 2535MHz	22.1	16GB GDDR6128-bit	288GB/s	165 W	$499
GeForce RTX 4090 D	December 28, 2023	AD102-250-A1	–	76B	608.5 mm²125.4 M/mm²	14,592	2280 / 2520MHz	73.5	24GB GDDR6X384-bit	1,008GB/s	425 W	$1,599
GeForce RTX 4070 Super	January 17, 2024	AD104-350-A1	TSMC 4 nm	36B	294.5 mm²121.6 M/mm²	7,168	1980 / 2475MHz	35.5	12GB GDDR6X192-bit	504GB/s	220 W	$599
GeForce RTX 4070 Ti Super	January 24, 2024	AD103-275-A1	TSMC 4 nm	46B	378.6 mm²121.2 M/mm²	8,448	2340 / 2610MHz	44.1	16GB GDDR6X256-bit	672GB/s	285 W	$799
GeForce RTX 4080 Super	January 31, 2024	AD103-400-A1	TSMC 4 nm	46B	378.6 mm²121.2 M/mm²	10,240	2295 / 2505MHz	52.2	16GB GDDR6X256-bit	736GB/s	320 W	$999
GeForce RTX 4070 (AD103)	March 2024	AD103-175-K1-A1	–	46B	378.6 mm²121.2 M/mm²	5,888	1920 / 2475MHz	29.1	12GB GDDR6X192-bit	504.2GB/s	200 W	$599
GeForce RTX 4060 (AD106)	April 2024	AD106-155-A1	–	23B	187.8 mm²121.9 M/mm²	3,072	1830 / 2460MHz	15.1	8GB GDDR6128-bit	272GB/s	115 W	$299
GeForce RTX 4060 Ti (AD104)	April 2024	AD104-150-K1-A1	–	36B	294.5 mm²121.6 M/mm²	4,352	2310 / 2535MHz	22.1	8GB GDDR6128-bit	288GB/s	160 W	$399
GeForce RTX 4070 Ti Super (AD102)	June 10, 2024	AD102-175-KEF-A1	–	76B	608.5 mm²125.4 M/mm²	8,448	2340 / 2610MHz	44.1	16GB GDDR6X256-bit	672.3GB/s	285 W	$799
GeForce RTX 4070 GDDR6	August 20, 2024	AD104-2??-A1	–	36B	294.5 mm²121.6 M/mm²	5,888	1920 / 2475MHz	29.1	12GB GDDR6192-bit	480GB/s	200 W	$599
Workstation 7
RTX 6000 Ada Generation	Dec 3, 2022	AD102-870-A1	TSMC 4 nm	76B	608.5 mm²125.4 M/mm²	18,176	915 / 2505MHz	91.1	48GB GDDR6 ECC384-bit	960GB/s	300 W	$6,799
RTX 4000 SFF Ada Generation	Mar 21, 2023	AD104	TSMC 4 nm	36B	294.5 mm²121.6 M/mm²	6,144	720 / 1560MHz	19.2	20GB GDDR6160-bit	280GB/s	70 W	$1,250
RTX 4000 Ada Generation	Aug 9, 2023	AD104	TSMC 4 nm	36B	294.5 mm²121.6 M/mm²	6,144	1500 / 2175MHz	26.7	20GB GDDR6160-bit	360GB/s	130 W	$1,250
RTX 4500 Ada Generation	Aug 9, 2023	AD103	TSMC 4 nm	46B	378.6 mm²121.2 M/mm²	7,680	2070 / 2580MHz	39.6	24GB GDDR6 ECC192-bit	432GB/s	210 W	$2,250
RTX 5000 Ada Generation	Aug 9, 2023	AD102-850- KAB-A1	TSMC 4 nm	76B	608.5 mm²125.4 M/mm²	12,800	1155 / 2550MHz	65.3	32GB GDDR6 ECC256-bit	576GB/s	250 W	$4,000
RTX 5880 Ada Generation	Jan 5, 2024	AD102	TSMC 4 nm	76B	608.5 mm²125.4 M/mm²	14,080	975 / 2460MHz	69.3	48GB GDDR6 ECC384-bit	960GB/s	285 W	$6,999
RTX 2000 Ada Generation	Feb 12, 2024	AD107-875-A1	TSMC 4 nm	19B	158.7 mm²119.1 M/mm²	2,816	1620 / 2130MHz	12.0	16GB GDDR6128-bit	224GB/s	70 W	$649
Datacenter 3
L40	October 13, 2022	AD102	TSMC 4 nm	76B	608.5 mm²125.4 M/mm²	18,176	735 / 2490MHz	90.5	48GB GDDR6384-bit	864GB/s	300 W	$7,000
L4	March 21, 2023	AD104	TSMC 4 nm	36B	294.5 mm²121.6 M/mm²	7,424	795 / 2040MHz	30.3	24GB GDDR6192-bit	300GB/s	72 W	$2,500
L40S	August 8, 2023	AD102	TSMC 4 nm	76B	608.4 mm²125.4 M/mm²	18,176	1065 / 2520MHz	91.6	48GB GDDR6384-bit	864GB/s	350 W	$7,500

Blackwell

2024–2025 · TSMC 4 nm · 21 GPUs · 16 benchmarked · up to 208B transistors

Blackwell joins two reticle-sized dies into one 208-billion-transistor B200 package for the datacenter, with FP4 tensor maths. The single-die GB202 to GB207 power GeForce RTX 50 and RTX PRO cards from 2025, the first with GDDR7 memory.

GPU	Launch	Die	Node	Transistors	Die size	Shaders	Clock	FP32 TFLOPS	Memory	Bandwidth	TDP	Launch price
Desktop 10
GeForce RTX 5080	January 30, 2025	GB203-400-A1	TSMC 4 nm	46B	378 mm²120.6 M/mm²	10,752	2295 / 2617MHz	56.3	16GB GDDR7256-bit	960GB/s	360 W	$999
GeForce RTX 5090	January 30, 2025	GB202-300	TSMC 4 nm	92B	750 mm²122.9 M/mm²	21,760	2017 / 2407MHz	105	32GB GDDR7512-bit	1,792GB/s	575 W	$1,999
GeForce RTX 5090 D	January 30, 2025	GB202-250-A1	–	92B	750 mm²122.9 M/mm²	21,760	2017 / 2407MHz	105	32GB GDDR7512-bit	1,792GB/s	575 W	$2,299
GeForce RTX 5070 Ti	February 20, 2025	GB203-300-A1	TSMC 4 nm	46B	378 mm²120.6 M/mm²	8,960	2300 / 2452MHz	43.9	16GB GDDR7256-bit	896GB/s	300 W	$749
GeForce RTX 5070	March 4, 2025	GB205-300-A1	TSMC 4 nm	31B	263 mm²118.3 M/mm²	6,144	2325 / 2512MHz	30.7	12GB GDDR7192-bit	672GB/s	250 W	$549
GeForce RTX 5060 Ti	April 16, 2025	GB206-300-A1	–	22B	181 mm²121 M/mm²	4,608	2407 / 2572MHz	23.7	8GB GDDR7128-bit	448GB/s	180 W	$379
GeForce RTX 5060 Ti 16GB	April 16, 2025	GB206-300-A1	TSMC 4 nm	22B	181 mm²121 M/mm²	4,608	2407 / 2572MHz	23.7	16GB GDDR7128-bit	448GB/s	180 W	$429
GeForce RTX 5060	May 19, 2025	GB206-250-A1	TSMC 4 nm	22B	181 mm²121 M/mm²	3,840	2280 / 2497MHz	19.2	8GB GDDR7128-bit	448GB/s	145 W	$299
GeForce RTX 5050	July 1, 2025	GB207-300-A1	TSMC 4 nm	17B	149 mm²113.4 M/mm²	2,560	2317 / 2570MHz	13.2	8GB GDDR6128-bit	320GB/s	130 W	$249
GeForce RTX 5090 D V2	August 15, 2025	GB202-240-K1-A1	–	92B	750 mm²122.9 M/mm²	21,760	2017 / 2407MHz	105	24GB GDDR7384-bit	1,344GB/s	575 W	$2,299
Workstation 7
RTX PRO 4000 Blackwell	Mar 18, 2025	GB203	TSMC 4 nm	46B	378 mm²120.6 M/mm²	8,960	1590 / 2617MHz	46.9	24GB GDDR7 ECC192-bit	672GB/s	140 W	$1,546
RTX PRO 4500 Blackwell	Mar 18, 2025	GB203	TSMC 4 nm	46B	378 mm²120.6 M/mm²	10,496	1635 / 2407MHz	54.9	32GB GDDR7 ECC256-bit	896GB/s	200 W	$2,623
RTX PRO 5000 Blackwell	Mar 18, 2025	GB202	TSMC 4 nm	92B	750 mm²122.9 M/mm²	14,080	1740 / 2377MHz	73.7	48GB GDDR7 ECC384-bit	1,344GB/s	300 W	$4,569
RTX PRO 6000 Blackwell Max-Q Workstation Edition	Mar 18, 2025	GB202	TSMC 4 nm	92B	750 mm²122.9 M/mm²	24,064	900 / 2617MHz	126	96GB GDDR7 ECC512-bit	1,792GB/s	300 W	$8,565
RTX PRO 6000 Blackwell Workstation Edition	Mar 18, 2025	GB202	TSMC 4 nm	92B	750 mm²122.9 M/mm²	24,064	1590 / 2617MHz	126	96GB GDDR7 ECC512-bit	1,792GB/s	600 W	$8,565
RTX PRO 2000 Blackwell	Aug 11, 2025	GB206	TSMC 4 nm	22B	181 mm²121 M/mm²	4,352	982 / 1957MHz	17.0	16GB GDDR7128-bit	288GB/s	70 W	$722
RTX PRO 4000 Blackwell SFF Edition	Aug 11, 2025	GB203	TSMC 4 nm	46B	378 mm²120.6 M/mm²	8,960	1342MHz	24.1	24GB GDDR7192-bit	432GB/s	70 W	–
Datacenter 4
B100	March 18, 2024	GB102	TSMC 4 nm	208B	–	–	–	–	192GB HBM3e8192-bit	8,000GB/s	700 W	$32,500
B200	November 2024	GB100	TSMC 4 nm	208B	–	18,944	–	74.5	192GB HBM3e8192-bit	8,000GB/s	1000 W	$40,000
RTX PRO 6000 Blackwell Server Edition	May 2025	GB202	TSMC 4 nm	92B	750 mm²122.9 M/mm²	24,064	1590 / 2610MHz	126	96GB GDDR7 ECC512-bit	1,792GB/s	600 W	$8,565
B300	Sep 11, 2025	GB110	TSMC 4 nm	208B	–	18,944	–	77.0	288GB HBM3e8192-bit	8,000GB/s	1400 W	$40,000

How to read this

Transistor count is how much logic the designers could afford; die size is what the chip costs to make, because wafers are priced by area and defects scale with it. Density (millions of transistors per mm², under each die size) shows what a process node delivered. Cards that share a die code share silicon: the difference between them is how much of it is switched on, how fast it runs and what memory hangs off it. FP32 TFLOPS is the peak at boost clock, shaders × 2 × clock; RDNA 3 and RDNA 4 figures count dual-issue. A faded picture is a representative card of that architecture, not the exact model.

Specifications from Wikipedia’s NVIDIA, AMD and Intel Arc lists, available under CC BY-SA 4.0, checked against vendor launch specifications, with corrections where the two disagree. Cards we benchmarked use our own verified specs. Benchmark results on the linked pages are our own.