Fetchy

New GeForce RTX 4080 Super vs GeForce RTX 4070.

GeForce RTX 4080 Super, 16 GB
€1,109last new priceNot sold new
vs
GeForce RTX 4070, 12 GB
Not enough prices yet

The GeForce RTX 4080 Super is no longer sold new

Miso compared the prices

Without a current new price there is no value score, so there is no winner on value.

Compare on Used

GeForce RTX 4080 Super

GeForce RTX 4080 Super is no longer sold new, so it has no value score. The price is the maker's last list price.

How the value score works
New priceNot sold new€1,109The maker's last list price

GeForce RTX 4070

Not enough priced models on this basis yet

How the value score works
New price-Not enough prices yet

Where they sit

Not enough priced models on this basis to map yet

Where they differ

Why the GeForce RTX 4080 Super is better

  • +23 TFLOPSFloating-point performance, 52.2 TFLOPS vs 29.2 TFLOPS
  • +7,359PassMark (G3D) result, 34,211 vs 26,852
  • +4,352 shading unitsShading units, 10,240 vs 5,888
  • +48 ROPsRender output units (ROPs), 112 vs 64
  • +127.2 GPixel/sPixel rate, 285.6 GPixel/s vs 158.4 GPixel/s

Why the GeForce RTX 4070 is better

  • 60 mm shorterLength, 244 mm vs 304 mm
  • 120 W less powerThermal Design Power (TDP), 200 W vs 320 W
See all differences

Section scores

Each section scores 0 to 100 among all graphics cards. The dashed line is their average.

How sections are scored
Hover a name or the outline to preview. Click to switch.
graphics card average
Strong performance and features
Section scores and category average
Section scoresGeForce RTX 4080 SuperGeForce RTX 4070Category average
Performance573432
Memory402628
Efficiency406635
Features505034

Full specs

GeForce RTX 4080 Super

GeForce RTX 4070

Performance

11 specs

GPU clock speed
2,290 MHz Better
1,920 MHz
GPU turbo
2,550 MHz Better
2,480 MHz
Pixel rate
285.6 GPixel/s Better
158.4 GPixel/s
Show all specs
Texture rate
816 GTexels/s Better
456.3 GTexels/s
Floating-point performance
52.2 TFLOPS Better
29.2 TFLOPS
Shading units
10,240 Better
5,888
Texture mapping units (TMUs)
320 Better
184
Render output units (ROPs)
112 Better
64
Tensor cores
320 Better
184
Ray tracing cores
80 Better
46
PassMark (G3D) result
34,211 Better
26,852

Memory

5 specs

VRAM
16 GB Better
12 GB
Maximum memory bandwidth
736 GB/s Better
504.2 GB/s
Effective memory speed
23,000 MHz Better
21,000 MHz
Show all specs
Memory bus width
256-bit Better
192-bit
GDDR version
GDDR6X
GDDR6X

Features

7 specs

Supports ray tracing
Yes
Yes
Supports DLSS
Yes
Yes
DLSS version
3.5
3.5
Show all specs
DirectX version
DirectX 12 Ultimate
DirectX 12 Ultimate
OpenGL version
4.6
4.6
AMD SAM / Intel Resizable BAR
Resizable BAR
Resizable BAR
Supported displays
4
4

Ports

4 specs

DisplayPort outputs
3
3
HDMI ports
1
1
Show all specs
HDMI version
HDMI 2.1a
HDMI 2.1a
DisplayPort version
DisplayPort 1.4a
DisplayPort 1.4a

General info

8 specs

Thermal Design Power (TDP)
320 W
200 W Better
Number of transistors
45,900 million Better
35,800 million
PCI Express (PCIe) version
4.0
4.0
Show all specs
Semiconductor size
4 nm
4 nm
Length
304 mm
244 mm Better
Power connector
16-pin (12VHPWR)
8-pin PCIe or 16-pin (12VHPWR)
GPU architecture
Ada Lovelace
Ada Lovelace
Release date
31 Jan 2024
13 Apr 2023
  • Opens the maker's spec page
  • Better value
  • Example marks a value with no maker source

Which one to pick

Pick the GeForce RTX 4080 Super if…

  • You want the newer model, from 2024.
  • You plan to keep it for longer.
  • You want the 16 GB version.
See the GeForce RTX 4080 Super

Pick the GeForce RTX 4070 if…

  • A model from 2023 does what you need.
  • You plan to upgrade again in a few years.
  • You want the 12 GB version.
See the GeForce RTX 4070
Coming soon: alerts for two models

Get a deal alert for either one.

One alert watches the GeForce RTX 4080 Super and GeForce RTX 4070 together. A deal alert costs 1 credit per 30 days.

We keep your email and this topic to send you a confirmation and price news about it. Unsubscribe any time. Privacy policy

See pricing