Skip to content
RU

Graphics cards

235 models. Pick two and compare.

Benchmark rankings

NVIDIA · Blackwell

GeForce RTX 5060 Ti

  • 180 W
Render (OptiX) · 22 of 6729%

NVIDIA · Ada Lovelace

GeForce RTX 4060 Ti

Render (OptiX) · 24 of 6726%

NVIDIA · Blackwell

GeForce RTX 5060

  • 8 GB GDDR7
  • 145 W
  • PSU 550 W+
Render (OptiX) · 25 of 6725%
At launch
$299

AMD · RDNA 3

Radeon RX 7900 XT

  • 20 GB GDDR6
  • 315 W
  • PSU 750 W+
Render (HIP) · 2 of 1384%
At launch
$899

NVIDIA · Ada Lovelace

GeForce RTX 4070 Laptop GPU

  • 8 GB GDDR6
  • 35–115 W
Render (OptiX) · 26 of 6724%

NVIDIA · Ampere

GeForce RTX 3070 Ti

  • 8 GB GDDR6X
  • 290 W
  • PSU 750 W+
Render (OptiX) · 27 of 6723%
At launch
$599

NVIDIA · Blackwell

GeForce RTX 5060 Laptop GPU

  • 8 GB GDDR7
  • 45–100 W
Render (OptiX) · 28 of 6723%

NVIDIA · Ada Lovelace

GeForce RTX 4060 Laptop GPU

  • 8 GB GDDR6
  • 35–115 W
Render (OptiX) · 29 of 6722%

AMD · RDNA 4

Radeon RX 9070 XT

  • 16 GB GDDR6
  • 304 W
  • PSU 750 W+
Render (HIP) · 3 of 1376%
At launch
$599

NVIDIA · Ampere

GeForce RTX 3070

  • 8 GB GDDR6
  • 220 W
  • PSU 650 W+
Render (OptiX) · 30 of 6721%
At launch
$499

NVIDIA · Ada Lovelace

GeForce RTX 4060

  • 8 GB GDDR6
  • 115 W
  • PSU 550 W+
Render (OptiX) · 31 of 6721%
At launch
$299

NVIDIA · Blackwell

GeForce RTX 5050 Laptop GPU

  • 8 GB GDDR7
  • 35–100 W
Render (OptiX) · 33 of 6720%

AMD · RDNA 3

Radeon RX 7900 GRE

  • 16 GB GDDR6
  • 260 W
Render (HIP) · 4 of 1367%
At launch
$549

NVIDIA · Blackwell

GeForce RTX 5050

  • 8 GB GDDR6
  • 130 W
  • PSU 550 W+
Render (OptiX) · 36 of 6719%
At launch
$249

NVIDIA · Turing

GeForce RTX 2080 Ti

  • 11 GB GDDR6
  • 250 W
Render (OptiX) · 37 of 6719%
At launch
$999

AMD · RDNA 4

Radeon RX 9070

  • 16 GB GDDR6
  • 220 W
  • PSU 650 W+
Render (HIP) · 5 of 1365%
At launch
$549

NVIDIA · Ampere

GeForce RTX 3060 Ti

  • 8 GB GDDR6 / GDDR6X
  • 200 W
  • PSU 600 W+
Render (OptiX) · 38 of 6718%
At launch
$399

The bar on a card is the result in Blender Benchmark 4.5.0 relative to the leader on the same render backend (OptiX, HIP, oneAPI or Metal; rendering, not games); “N of M” is the position among M tested models. Manufacturer's suggested launch price (US), rounded to whole dollars.