Nvidia RTX 5070 Ti vs RTX 4080 Super: Which is best for gaming in 2025?

The Nvidia RTX 5070 Ti and RTX 4080 Super are some of the most powerful gaming GPUs in the market. Shipping with 16 GB video memory, these cards represent the upper quartile of enthusiast-grade GPUs with unparalleled 4K capabilities. They are great for workstations, given the sheer rendering prowess Nvidia ships them with. However, choosing between the last generation for some savings and the newer $750 offering can be tough, especially given the many similarities.

In this article, we will go over the specs and gaming performance details to decide which card represents a better investment.

---

The Nvidia RTX 5070 Ti and RTX 4080 Super are some of the fastest gaming GPUs

The Nvidia RTX 5070 Ti targets high-end 1440p and 4K gaming (Image via Gigabyte)

The Nvidia RTX 5070 Ti is based on the newer Blackwell architecture. Unlike Ada, the newer GPUs unify a few of the moving parts of the GPUs and improve AI capabilities with a dedicated management chip. This makes a more efficient design that draws significantly less power to deliver similar performance. Let's review the on-paper specs to get a better idea.

---

Specs comparison

The RTX 5070 Ti is based on the GB203 graphics chip, a 378 mm² design that's slightly cut down from the flagship GB202. The 4080 Super shares a similar story: a 379 mm² AD103 graphics chip that isn't as fast as the AD102. Moreover, you get 10,240 CUDA cores with the last-gen 80-class card compared to the 8,960 on the 5070 Ti. Similarly, RT cores and Tensor core counts are cut by 13% on the 5070 Ti, but note that each individual core is faster than the last generation.

The Blackwell architecture's design changes, coupled with the stricter hardware budget, make it a 300W GPU, down from the 320W 4080 Super. Here's a detailed specs comparison table:

Spec	RTX 5070 Ti	RTX 4080 SUPER
Architecture / Process	Blackwell 2.0 / TSMC 5 nm	Ada Lovelace / TSMC 4N
CUDA Cores	8,960	10,240
Base Clock	~ 2,295 MHz	~ 2,295 MHz
Boost Clock	~ 2,452 MHz	~ 2,550 MHz
Memory Size / Type	16 GB GDDR7	16 GB GDDR6X
Memory Bus / Interface Width	256-bit	256-bit
Effective Memory Speed / Data Rate	~ 28 Gbps	~ 23 Gbps
Memory Bandwidth	~ 896 GB/s	~ 736.3 GB/s
ROPs (Raster Ops)	96	112
Texture Units (TMUs)	280	320
TGP / Power Consumption	~ 300 W	~ 320 W

The 4080 Super was introduced at $999, with the 5070 Ti targeting a $749 price tag. While the latter maintains the price, with a few models currently selling for $750, that isn't the case for the 4080 Super, as it has ended production. If you aren't getting a good deal or already own the GPU, paying the premium may not be worthwhile.

---

Performance comparison

The Nvidia RTX 4080 Super continues to be a capable 4K gaming GPU (Image via Gigabyte)

Here's how the Nvidia RTX 5070 Ti and RTX 4080 Super compare in some of the latest video games. The GPUs were tested at 4K with the Ryzen 7 9800X3D. The numbers are from the YouTube channel Testing Games.

	Nvidia RTX 5070 Ti	Nvidia RTX 4080 Super
Microsoft Flight Simulator 2024	46 FPS	47 FPS
Black Myth: Wukong	53 FPS	52 FPS
S.T.A.L.K.E.R. 2	49 FPS	49 FPS
Cyberpunk 2077	57 FPS	58 FPS
God of War: Ragnarök	94 FPS	92 FPS
Kingdom Come Deliverance 2	54 FPS	54 FPS
Horizon Forbidden West	71 FPS	72 FPS
Forza Horizon 5	117 FPS	116 FPS
Ghost of Tsushima	55 FPS	57 FPS
Silent Hill 2	69 FPS	69 FPS

The RTX 4080 Super is faster, but only marginally (approximately 1.3%). They're within 1-4 FPS of each other, lying between the error margin. This means native capabilities would largely be the same across the cards. With DLSS frame generation turned on, the 5070 Ti would take a solid lead in supported gains, as it supports 2x, 3x, and 4x presets missing on the older GPU.

Given that the 5070 Ti costs less, comes with better future proofing, and bundles the same 16 GB VRAM buffer, we recommend buying the newer graphics card. This gives you better peace of mind and your bank a break.