NVIDIA GeForce GTX 560 Ti vs NVIDIA GeForce GTX 670 Comparison

Head-to-Head Benchmarks

The recorded data includes a single directly comparable benchmark between the GeForce GTX 670 and the GeForce GTX 560 Ti: Geekbench OpenCL. In this test, the GTX 670 scores 15,341 against 10,690 for the GTX 560 Ti, a decisive 43.5% advantage. This is not a marginal lead; it is a substantial gap that places the GTX 670 in a different performance tier entirely.

To contextualize the GTX 670's result, the database shows its average benchmark score across all tests is 12,773, which places it at the 52nd percentile of all GPUs. Its nearest rivals include the NVIDIA GeForce GTX 590 at 12,830 (0.4% higher), the AMD Radeon Pro 455 at 12,831 (0.4% higher), the AMD Radeon RX 7600M XT at 12,710 (0.5% lower), and the AMD FirePro W5100 at 12,847 (0.6% lower). The GTX 670 sits essentially in the middle of this tight cluster, with deltas ranging from -0.6% to +0.5%. This suggests the GTX 670's performance is very consistent with its immediate peers, making its 43.5% lead over the GTX 560 Ti all the more striking.

On the other side, the GTX 560 Ti's average benchmark score is 10,690, which lands it at the 49th percentile of all GPUs. Its nearest rivals include the AMD Radeon RX 6600S at 10,629 (0.6% lower), the NVIDIA Quadro K2200 at 10,761 (0.7% higher), the AMD Radeon R9 M275X at 10,582 (1% lower), and the AMD Radeon Pro 450 at 10,804 (1.1% higher). The GTX 560 Ti also sits well within its expected band, with deltas ranging from -1.1% to +1%. The data indicates the GTX 560 Ti is a mid-pack performer in its era, but the GTX 670's single benchmark result shows a decisive generational leap forward.

Architecture Differences

The GTX 670 is built on the GK104 chip using the Kepler architecture, fabricated on a 28 nm process at TSMC. In contrast, the GTX 560 Ti uses the GF114 chip with the Fermi 2.0 architecture, also from TSMC but on a 40 nm process. This process shrink is significant: the Kepler die is 294 mm² and contains 3,540 million transistors, while the Fermi die is larger at 332 mm² but holds only 1,950 million transistors. Consequently, transistor density more than doubles, from 5.9 million per mm² on the Fermi to 12.0 million per mm² on the Kepler. The data shows a fundamental shift in design efficiency between the two generations.

Clock behavior also differs sharply. The GTX 670 has a base clock of 915 MHz and a boost clock of 980 MHz, allowing dynamic frequency scaling under load. The GTX 560 Ti lists no base or boost clock in the database, only a memory clock of 1002 MHz (4 Gbps effective), compared to the GTX 670's memory clock of 1502 MHz (6 Gbps effective). The absence of a boost mechanism on the older card suggests a fixed-frequency design, whereas the newer card can push beyond its base speed when thermal headroom permits.

Memory architecture shows both cards use GDDR5 with a 256-bit bus, but the GTX 670 doubles capacity to 2 GB versus 1024 MB, and its bandwidth climbs from 128.3 GB/s to 192.3 GB/s. The shading unit count is a dramatic difference: the GTX 670 has 1,344 shading units against 384 for the GTX 560 Ti, a 3.5x increase. Texture mapping units also jump from 64 to 112, while both cards keep 32 ROPs. These raw counts explain why the pixel rate rises from 13.17 GPixel/s to 27.44 GPixel/s and the texture rate from 52.67 GTexel/s to 109.8 GTexel/s. FP32 throughput nearly doubles, from 1,263.4 GFLOPS to 2.634 TFLOPS. Both cards share a 170 W TDP, a 450 W suggested PSU, dual-slot width, and two 6-pin power connectors, so the performance gain comes without a higher power draw.

PCIe interface differs: the GTX 560 Ti uses PCIe 2.0 x16, while the GTX 670 uses PCIe 3.0 x16. Display outputs also change, with the GTX 670 offering 2x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2, while the GTX 560 Ti offers 2x DVI and 1x mini-HDMI 1.3a. API support shows the GTX 670 lists Vulkan 1.2.175, whereas the GTX 560 Ti has no Vulkan support; both support DirectX 12 (11_0) and OpenGL 4.6.

Where Each One Wins

There is only one recorded head-to-head metric, and the GTX 670 wins it outright. In Geekbench OpenCL, the GTX 670 leads by 43.5%, which is a dominant margin. The data therefore shows a one-sided picture: the GTX 670 is the clear winner in compute workloads that stress raw throughput. The GTX 560 Ti has no benchmark win in the database.

For gaming and rendering workloads, the architectural differences favor the GTX 670 resoundingly. The doubling of memory capacity from 1 GB to 2 GB, combined with a 50% increase in bandwidth, means the GTX 670 can handle higher resolutions and larger texture sets without stuttering. The boost clock of 980 MHz provides headroom that the GTX 560 Ti lacks entirely. The higher pixel rate (27.44 vs 13.17 GPixel/s) and texture rate (109.8 vs 52.67 GTexel/s) directly translate to faster fill-rate bound tasks. The 3.5x shading unit advantage makes the GTX 670 far more capable in geometry-heavy and compute-shader scenarios.

The GTX 560 Ti remains relevant only in niche scenarios where its 40 nm architecture is a known quantity, or where its smaller board length (229 mm versus 241 mm) fits a tighter case. The GTX 670 is longer by 12 mm, which could matter in compact builds, but that is the only dimension where the older card holds an edge. In every measurable compute and graphics metric the database records, the GTX 670 is superior.

Specification Differences

The database records the following differences between the two cards, with the GTX 670 first:

• Architecture: Kepler vs Fermi 2.0
• Process node: 28 nm vs 40 nm
• Transistors: 3,540 million vs 1,950 million
• Die size: 294 mm² vs 332 mm²
• Transistor density: 12.0M / mm² vs 5.9M / mm²
• Base clock: 915 MHz vs no base clock
• Boost clock: 980 MHz vs no boost clock
• Memory clock: 1502 MHz (6 Gbps effective) vs 1002 MHz (4 Gbps effective)
• Memory size: 2 GB vs 1024 MB
• Memory bandwidth: 192.3 GB/s vs 128.3 GB/s
• Shading units: 1344 vs 384
• TMUs: 112 vs 64
• Pixel rate: 27.44 GPixel/s vs 13.17 GPixel/s
• Texture rate: 109.8 GTexel/s vs 52.67 GTexel/s
• FP32: 2.634 TFLOPS vs 1,263.4 GFLOPS
• Bus interface: PCIe 3.0 x16 vs PCIe 2.0 x16
• Display outputs: 2x DVI, 1x HDMI 1.4a, 1x DisplayPort 1.2 vs 2x DVI, 1x mini-HDMI 1.3a
• Vulkan support: 1.2.175 vs none
• Dimensions: 241 mm length vs 229 mm length; height and width only recorded for the GTX 670 (111 mm height, 38 mm width)
• Release date: 2012-05-09 vs 2011-01-24

The two cards share identical TDP (170 W), power connectors (2x 6-pin), suggested PSU (450 W), GDDR5 memory type, 256-bit bus width, ROPs (32), DirectX support (12 (11_0)), OpenGL support (4.6), slot width (dual-slot), and production status (end-of-life). The GTX 670 has a launch MSRP of 399 USD; the GTX 560 Ti has a launch MSRP of 249 USD.

FAQ

Q: Which card has a higher average benchmark score?

A: The GTX 670 averages 12,773, while the GTX 560 Ti averages 10,690, a difference of 1,933 points.

Q: How much faster is the GTX 670 in the OpenCL test?

A: The GTX 670 scores 15,341 versus 10,690, a 43.5% lead.

Q: Does the GTX 670 support more display outputs?

A: Yes, the GTX 670 offers 2x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2, while the GTX 560 offers 2x DVI and 1x mini-HDMI 1.3a.

Q: Which card has a higher pixel fill rate?

A: The GTX 670 has a pixel rate of 27.44 GPixel/s, more than double the 13.17 GPixel/s of the GTX 560 Ti.

Q: Do both cards have the same power draw?

A: Yes, both are rated at 170 W TDP and use two 6-pin power connectors.

Q: Which card is older?

A: The GTX 560 Ti was released on 2011-01-24, while the GTX 670 was released on 2012-05-09.

The Verdict

The data is unambiguous. The NVIDIA GeForce GTX 670 is the superior card across every recorded metric. It holds a 43.5% lead in the only direct benchmark, and its architectural advantages (Kepler architecture, 28 nm process, higher transistor count, 3.5x shading units, double the memory, and 50% more bandwidth) translate to roughly double the pixel and texture rates. The GTX 670 also adds Vulkan support, newer PCIe interface, and a DisplayPort output, features the GTX 560 Ti lacks.

Users who already own a GTX 560 Ti and are considering an upgrade will see a massive gain in compute workloads and memory-intensive tasks. Those who are deciding between the two for a legacy build should pick the GTX 670 without hesitation, as it occupies a higher percentile (52 vs 49) and averages 19.5% higher scores. The only consideration against the GTX 670 is its longer board length (241 mm) and the higher launch MSRP (399 USD vs 249 USD), but the data shows the performance premium is justified. The GTX 560 Ti remains a functional card for light gaming, but the GTX 670 is the definitive choice for any workload that stresses GPU compute or memory bandwidth.