GPU Rankings

Auto-calculated rankings based on performance for 146 graphics cards

Filter by Manufacturer

Filter by Type

Search GPUs

Overall Rank	GPU	Release Date	VRAM	TFLOPS	Price	Performance Index
#1🥇	GeForce RTX 5090 NVIDIA	Jan 2025	32GB GDDR7	104.80 TF	$1999.00	100.00
#2🥈	GeForce RTX 5090D NVIDIA	Jan 2025	32GB GDDR7	104.80 TF	$1999.00	96.16
#3🥉	GeForce RTX 4090 NVIDIA	Oct 2022	24GB GDDR6X	82.58 TF	$1599.00	79.76
#4	GeForce RTX 5080 NVIDIA	Jan 2025	16GB GDDR7	56.28 TF	$999.00	72.70
#5	GeForce RTX 4080 SUPER NVIDIA	Jan 2024	16GB GDDR6X	52.22 TF	$999.00	65.68
#6	Radeon RX 7900 XTX AMD	Dec 2022	24GB GDDR6	61.44 TF	$949.00	63.47
#7	GeForce RTX 4080 NVIDIA	Nov 2022	16GB GDDR6X	48.74 TF	$1099.00	62.78
#8	GeForce RTX 3090 Ti NVIDIA	Mar 2022	24GB GDDR6X	40.00 TF	$900.00	60.41
#9	GeForce RTX 5070 Ti NVIDIA	Feb 2025	16GB GDDR7	44.00 TF	$899.00	59.72
#10	Radeon RX 9070 XT AMD	Mar 2025	16GB GDDR6	48.70 TF	$599.00	56.93
#11	Radeon RX 7900 XT AMD	Dec 2022	20GB GDDR6	51.48 TF	$799.00	56.66
#12	GeForce RTX 3080 Ti NVIDIA	Jun 2021	12GB GDDR6X	34.10 TF	$600.00	55.86
#13	GeForce RTX 4070 Ti SUPER NVIDIA	Jan 2024	16GB GDDR6X	44.10 TF	$799.00	55.72
#14	GeForce RTX 4070 Ti NVIDIA	Jan 2023	12GB GDDR6X	40.09 TF	$799.00	55.02
#15	GeForce RTX 3090 NVIDIA	Sep 2020	24GB GDDR6X	35.58 TF	$800.00	54.69
#16	GeForce RTX 3080 12GB NVIDIA	Jan 2022	12GB GDDR6X	30.64 TF	$525.00	54.30
#17	Radeon RX 9070 AMD	Mar 2025	16GB GDDR6	36.10 TF	$549.00	53.46
#18	GeForce RTX 4090 Laptop GPU NVIDIA	Jan 2023	16GB GDDR6	32.98 TF	N/A	53.10
#19	Radeon RX 7900 GRE AMD	Sep 2023	16GB GDDR6	45.98 TF	$549.00	52.65
#20	GeForce RTX 4070 Super NVIDIA	Jan 2024	12GB GDDR6X	35.48 TF	$599.00	50.62
#21	GeForce RTX 4090 Laptop NVIDIA	Feb 2023	16GB GDDR6	39.68 TF	N/A	50.45
#22	Radeon RX 6900 XT AMD	Dec 2020	16GB GDDR6	23.04 TF	$350.00	49.59
#23	GeForce RTX 3080 NVIDIA	Sep 2020	10GB GDDR6X	29.77 TF	$475.00	48.65
#24	GeForce RTX 4070 NVIDIA	Apr 2023	12GB GDDR6X	29.15 TF	$499.00	47.93
#25	GeForce RTX 5070 NVIDIA	Feb 2025	12GB GDDR7	30.87 TF	$549.00	46.91

Showing 1 to 25 of 146 GPUs

Page 1 of 6

How GPU Rankings Work

Overall Rank: Combined score based on performance, value, and efficiency
🥇🥈🥉: Top 3 performers in overall ranking

📖 Technical Terms Explained

Performance & Specifications

Performance Index

A normalized benchmark score used to rank and compare hardware across different generations and architectures. For GPUs, combines gaming performance at multiple resolutions with ray tracing capabilities. Higher numbers always indicate better overall performance. This single metric allows fair comparison between vastly different products - comparing a 2020 GPU to a 2024 GPU becomes straightforward.

Overall Rank

The position of this GPU in the complete performance rankings across all products in the database. Lower numbers indicate higher performance. Rank #1 is the fastest product available. Rankings update as new products release and benchmarks improve. Useful for quickly understanding where a product sits in the performance hierarchy - a Rank #15 GPU is high-end, while Rank #150 is budget-tier.

Release Date

The official launch date when the GPU became available for purchase. Note that availability may have been limited at launch due to supply constraints, especially during 2020-2022. Release date helps understand product age and whether newer alternatives exist. Generally, newer products offer better performance-per-dollar and modern features, though older products may provide excellent value on secondary market.

Price

Current retail pricing where available, typically showing MSRP or recent market pricing. Actual street prices vary by region, retailer, and availability. Used to calculate value rankings and price-to-performance ratios. Note: Prices fluctuate significantly - high-end GPUs may see $100+ swings, while budget products remain more stable. Check current retailer pricing before purchasing decisions.

TFLOPS (Teraflops)

Measures a GPU's theoretical floating-point operations per second - essentially raw mathematical compute power. While higher TFLOPS suggests more processing capability, it doesn't directly translate to gaming performance due to differences in architecture efficiency, memory bandwidth, and driver optimization. More useful for comparing GPUs within the same architecture family. For example, a 20 TFLOPS RDNA 3 GPU may outperform a 30 TFLOPS older architecture in actual games.

Filters & Categories

Form Factor

The physical design and target platform: Desktop (standard PCIe cards for tower PCs, maximum cooling and performance), Mobile (laptop GPUs with configurable power 35-175W), Gaming Console (custom GPUs in PlayStation/Xbox). Note: Mobile GPUs perform 30-60% slower than desktop counterparts due to power/thermal constraints.

Manufacturer

The company designing and producing the GPU. GPUs: NVIDIA, AMD, Intel (Arc series). Console GPUs use custom AMD designs. Manufacturer choice affects platform compatibility, feature support, and software optimization.

Include Workstation

Option to include professional workstation GPUs (NVIDIA RTX A-series, AMD Radeon Pro) in rankings alongside gaming GPUs. Workstation cards offer certified drivers, ECC memory, and enhanced reliability for professional applications but cost significantly more than gaming equivalents with similar raw performance. Typically excluded from gaming-focused comparisons.

Technologies

PCIe Generation

The PCI Express version supported for connecting to the motherboard. Each generation doubles bandwidth: PCIe 3.0 (1 GB/s per lane, 16 GB/s for x16) - older standard, PCIe 4.0 (2 GB/s per lane, 32 GB/s for x16) - current mainstream, PCIe 5.0 (4 GB/s per lane, 64 GB/s for x16) - latest high-end. Higher generations benefit fast NVMe SSDs and future high-bandwidth GPUs. Backward compatible.

Ray Tracing Support

Hardware acceleration for realistic lighting, reflections, and shadows in games. All modern GPUs include dedicated ray tracing hardware: NVIDIA (RT Cores, 3rd gen in RTX 40), AMD (Ray Accelerators, RDNA 2+), Intel (Ray Tracing Units, Arc). Ray tracing improves visual quality but reduces performance 20-50%. Mitigated by DLSS/FSR upscaling. Essential for modern AAA games with advanced lighting.

AI Upscaling (DLSS / FSR / XeSS)

Technologies rendering games at lower resolution then using AI to upscale, improving performance while maintaining quality. DLSS (NVIDIA RTX) uses Tensor cores for best image quality. FSR (AMD) works on any GPU through software. XeSS (Intel Arc) uses XMX engines. Typically provides 40-100% FPS improvement. Essential for 4K gaming and ray tracing viability.

Architecture

Architecture Generation

The microarchitecture design and manufacturing generation. Newer architectures generally offer better performance per watt and modern features. NVIDIA: Pascal (GTX 10) → Turing (RTX 20) → Ampere (RTX 30) → Ada (RTX 40) → Blackwell (RTX 50). AMD: Polaris → Vega → RDNA → RDNA 2 (RX 6000) → RDNA 3 (RX 7000) → RDNA 4 (RX 9000). Intel: Alchemist (Arc A-series).

VRAM Capacity

Amount of dedicated video memory for storing textures, models, and frame buffers. Requirements scale with resolution: 1080p = 6-8GB sufficient, 1440p = 8-12GB recommended, 4K = 12-16GB+ for maximum quality, Content Creation = 16-24GB for professional work. Insufficient VRAM causes stuttering, texture pop-in, and reduced quality.

TDP/TGP Range

Power consumption range indicating performance tier and cooling requirements. 75W (budget, no power connector), 100-150W (mainstream gaming), 200-300W (high-end), 350W+ (enthusiast flagship, requires quality PSU). Higher TDP/TGP generally indicates higher performance but demands better cooling and power delivery.