NVIDIA GeForce GTX 1660 SUPER vs NVIDIA GeForce GTX 950 Comparison

The GeForce GTX 950 and GeForce GTX 1660 SUPER represent two distinct eras of NVIDIA’s mainstream lineup, separated by four years of architectural evolution. The GTX 950, built on the Maxwell 2.0 design with the GM206 chip, was a 2015 entry point for 1080p gaming. The GTX 1660 SUPER, using the Turing-based TU116, arrived in 2019 as a more modern, higher-performance option. The benchmark data reveals a generational chasm, with the GTX 1660 SUPER dominating in every measurable compute workload, yet the older card still holds a place in specific legacy scenarios. This analysis breaks down the numbers from the FACT PACK to show exactly where each GPU stands.

Head-to-Head Benchmarks

The direct comparison in the FACT PACK is limited to two compute-oriented tests, and the results are lopsided. In Geekbench OpenCL, the GTX 950 scores 15,662 points, while the GTX 1660 SUPER delivers 52,490 points. That is a delta of -70.2%, meaning the GTX 1660 SUPER is roughly 3.35 times faster in this workload. The gap is nearly identical in Geekbench Vulkan, where the GTX 950 posts 16,734 points against the GTX 1660 SUPER’s 57,102 points, a -70.7% delta. These aren’t marginal improvements; they represent a complete generational overhaul in raw throughput.

Digging into the architectural specs behind those scores, the reasons become clear. The GTX 1660 SUPER has 1,408 shading units compared to the GTX 950’s 768, and its texture units number 88 versus 48. The newer card’s boost clock of 1785 MHz far exceeds the GTX 950’s 1188 MHz. This combination yields an FP32 performance of 5.027 TFLOPS for the GTX 1660 SUPER, against just 1.825 TFLOPS for the GTX 950. The pixel rate tells a similar story: 85.68 GPixel/s versus 38.02 GPixel/s, and the texture rate jumps from 57.02 GTexel/s to 157.1 GTexel/s. Every computational metric in the FACT PACK favors the newer card by a wide margin.

Memory bandwidth is another decisive factor. The GTX 950 uses 2 GB of GDDR5 on a 128-bit bus, yielding 105.8 GB/s. The GTX 1660 SUPER has 6 GB of GDDR6 on a 192-bit bus, achieving 336.0 GB/s. That’s more than triple the bandwidth, which directly impacts fill-rate-heavy scenes and compute tasks. The effective memory clock also nearly doubles, from 6.6 Gbps to 14 Gbps. While the head-to-head benchmarks don’t include a gaming test, the bandwidth and shader advantages strongly suggest the GTX 1660 SUPER would dominate in modern titles.

The GTX 1660 SUPER’s average benchmark score of 12,986 sits just 0.4% ahead of the GTX 950’s 13,189 in the broader database. However, this aggregate figure masks the fact that the newer card has a much wider benchmark suite, including PassMark tests for DirectX 9 through 12, where it scores 189, 65, 104, and 50 respectively, plus a G3D score of 12,699 and a GPU compute score of 5,076. The GTX 950 only has Geekbench Metal, OpenCL, and Vulkan results. In those shared tests, the GTX 1660 SUPER wins outright. The average score similarity is misleading because the test sets differ; the newer card’s Vulkan score alone is 3.4 times higher than the older card’s.

Where Each One Wins

The GTX 1660 SUPER wins in every category where data exists. Its Geekbench OpenCL score of 52,490 is 3.35 times the GTX 950’s 15,662, making it the clear choice for GPU-accelerated compute, such as video encoding, physics simulations, or machine learning inference. Its Vulkan score of 57,102 versus 16,734 indicates superior performance in modern graphics APIs, which are increasingly common in current game engines. The newer card also offers 6 GB of VRAM, which is essential for high-resolution textures and larger datasets, whereas the GTX 950’s 2 GB is a hard limitation for modern workloads.

The GTX 950 does have one niche advantage: power consumption. Its TDP is 90 W with a suggested PSU of 250 W, while the GTX 1660 SUPER draws 125 W and recommends a 300 W PSU. For a legacy system with an older power supply, the GTX 950 is the safer drop-in upgrade. It also requires only a 1x 6-pin power connector, whereas the GTX 1660 SUPER needs a 1x 8-pin. In terms of physical size, the GTX 950 is shorter at 202 mm (8 inches) versus the GTX 1660 SUPER’s 229 mm (9 inches), making it easier to fit in compact cases. The GTX 950 also has three DisplayPort 1.2 outputs, while the GTX 1660 SUPER has one DisplayPort 1.4a, two more for multi-monitor setups, though the newer card supports a newer standard.

For gaming, the data only indirectly points to the winner. The GTX 1660 SUPER’s higher pixel rate (85.68 GPixel/s) and texture rate (157.1 GTexel/s) mean it can handle higher resolutions and more detailed geometry. Its FP16 performance of 10.05 TFLOPS (2:1) is a feature the GTX 950 lacks entirely, which is relevant for applications using mixed-precision compute. The GTX 1660 SUPER’s PassMark DirectX 12 score of 50 and DirectX 11 score of 104 are low absolute numbers, but they represent the card’s ability to run those APIs; the GTX 950 has no such scores in the FACT PACK, so no direct comparison is possible. In every benchmark where both cards appear, the GTX 1660 SUPER wins.

The Verdict

Choose the GTX 1660 SUPER if you need performance in modern compute or graphics workloads. The data is unambiguous: it is 70.2% faster in Geekbench OpenCL and 70.7% faster in Geekbench Vulkan than the GTX 950. Its 5.027 TFLOPS of FP32 power, 336.0 GB/s of memory bandwidth, and 6 GB of VRAM make it suitable for tasks that would cripple the older card. For gaming at 1080p or higher with modern API support, the GTX 1660 SUPER is the only sensible option from these two.

Choose the GTX 950 only if your constraints are power or physical size. Its 90 W TDP and 250 W suggested PSU requirement are significantly lower than the GTX 1660 SUPER’s 125 W and 300 W. Its 202 mm length fits where a 229 mm card won’t. It also uses a 1x 6-pin connector, which is more common in older power supplies. However, this comes at a massive performance cost. The GTX 950’s 1.825 TFLOPS and 105.8 GB/s bandwidth are less than half the newer card’s figures. In any workload that stresses the GPU, the GTX 1660 SUPER is the clear winner. The average benchmark scores are close (13,189 vs 12,986), but that’s due to different test sets; in shared tests, the newer card wins by 70% or more. The GTX 950 is a legacy part for legacy systems, while the GTX 1660 SUPER is a capable modern GPU.

FAQ

Q: Which GPU has a higher Geekbench Vulkan score?

A: The NVIDIA GeForce GTX 1660 SUPER scores 57,102, while the GeForce GTX 950 scores 16,734, making the newer card 70.7% faster.

Q: How much memory bandwidth does each card have?

A: The GTX 950 has 105.8 GB/s from 2 GB of GDDR5 on a 128-bit bus. The GTX 1660 SUPER has 336.0 GB/s from 6 GB of GDDR6 on a 192-bit bus.

Q: What is the TDP difference between the two cards?

A: The GTX 950 has a TDP of 90 W with a suggested PSU of 250 W, while the GTX 1660 SUPER has a TDP of 125 W and a suggested PSU of 300 W.

Q: Are there any benchmarks where the GTX 950 wins?

A: No. In the head-to-head benchmarks (Geekbench OpenCL and Vulkan), the GTX 1660 SUPER wins both. The GTX 950 has 0 wins in the head-to-head data.

Q: What is the FP32 performance of each card?

A: The GTX 950 delivers 1.825 TFLOPS, while the GTX 1660 SUPER delivers 5.027 TFLOPS, more than 2.7 times higher.

Q: What are the power connector requirements?

A: The GTX 950 uses a single 6-pin connector, while the GTX 1660 SUPER requires a single 8-pin connector.

Architecture Differences

The two GPUs are built on fundamentally different architectures and process nodes. The GTX 950 uses the Maxwell 2.0 architecture (chip GM206) on a 28 nm process from TSMC, packing 2,940 million transistors into a 228 mm² die. This results in a transistor density of 12.9 million per mm². The GTX 1660 SUPER uses the Turing architecture (chip TU116) on a 12 nm process, also from TSMC, with 6,600 million transistors on a 284 mm² die, achieving a density of 23.2 million per mm². The newer process allows for more than double the transistor count in a slightly larger package.

Core configuration is a major differentiator. The GTX 950 has 768 shading units, 48 texture mapping units, and 32 ROPs. The GTX 1660 SUPER increases these to 1,408 shading units, 88 TMUs, and 48 ROPs. Clock speeds also rise significantly: the GTX 950’s base and boost are 1024 MHz and 1188 MHz, respectively, while the GTX 1660 SUPER runs at 1530 MHz base and 1785 MHz boost. The newer card also supports FP16 compute at 10.05 TFLOPS (2:1), a feature absent from the GTX 950’s spec sheet.

Memory subsystems differ completely. The GTX 950 uses 2 GB of GDDR5 at 6.6 Gbps effective on a 128-bit bus, yielding 105.8 GB/s. The GTX 1660 SUPER uses 6 GB of GDDR6 at 14 Gbps effective on a 192-bit bus, yielding 336.0 GB/s. Both cards support PCIe 3.0 x16, and both have DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4 API support. Display outputs differ: the GTX 950 has 1x DVI, 1x HDMI 2.0, and 3x DisplayPort 1.2, while the GTX 1660 SUPER has 1x DVI, 1x HDMI 2.0, and 1x DisplayPort 1.4a. Physical dimensions also vary, with the GTX 950 at 202 mm (8 inches) and the GTX 1660 SUPER at 229 mm (9 inches) long, the latter also having specified height (111 mm, 4.4 inches) and width (35 mm, 1.4 inches). Both are dual-slot cards, but the GTX 1660 SUPER’s larger die and higher TDP require more substantial cooling and power delivery.