NVIDIA GeForce GTX 1660 SUPER vs NVIDIA GeForce GTX 680 Comparison

Head-to-Head Benchmarks

The benchmark data shows a decisive victory for the NVIDIA GeForce GTX 1660 SUPER in every shared test. In Geekbench OpenCL, the GTX 1660 SUPER scores 52,490 against the GTX 680's 16,906, a difference of 67.8 percent in favor of the newer card. The Geekbench Vulkan result is even more lopsided: the GTX 1660 SUPER posts 57,102 while the GTX 680 manages 17,646, a 69.1 percent gap. These are not marginal improvements; the GTX 1660 SUPER more than triples the GTX 680's compute throughput in both APIs.

Looking at the broader database context, the GTX 680's average benchmark score of 14,150 places it in the 55th percentile of all GPUs, while the GTX 1660 SUPER's average of 12,986 sits in the 53rd percentile. This is an apparent contradiction: the newer card wins head-to-head by huge margins, yet its overall percentile is slightly lower. The explanation lies in the test suite composition. The GTX 680 has only three recorded benchmarks, all Geekbench compute tests, which tend to favor its architecture. The GTX 1660 SUPER has ten recorded benchmarks, including PassMark DirectX 9, 10, 11, and 12 tests, G2D, G3D, and GPU compute, which pull its average down. The GTX 1660 SUPER's PassMark G3D score of 12,699 and GPU compute score of 5,076 are strong, but its DirectX 9 score of 189 and DirectX 10 score of 65 are low, dragging the aggregate.

When comparing to nearest rivals, the GTX 680 sits within 1.4 percent of the AMD Radeon Vega 11 (14,352) and within 0.9 percent of the NVIDIA Tesla K10 (14,029). It is 0.9 percent behind the GTX 1070 Ti (14,277). The GTX 1660 SUPER's nearest rivals include the RTX 3050 Ti Mobile (12,940, 0.4 percent ahead), the RX 580 (12,928, 0.4 percent ahead), and the Tesla M2090 (13,075, 0.7 percent behind). These rivalries show that both cards are competitive within their respective performance tiers, but the direct comparison is unambiguous: the GTX 1660 SUPER dominates.

Architecture Differences

The two cards come from different architectural eras. The GTX 680 uses the GK104 chip built on the Kepler architecture, fabricated on a 28 nm process at TSMC. The GTX 1660 SUPER uses the TU116 chip on the Turing architecture, also from TSMC but on a 12 nm process. The transistor count tells the story of density: the GTX 680 has 3,540 million transistors on a 294 mm² die, giving a density of 12.0 million transistors per square millimeter. The GTX 1660 SUPER packs 6,600 million transistors on a slightly smaller 284 mm² die, achieving 23.2 million per square millimeter, nearly double the density.

Memory technology also diverges significantly. The GTX 680 uses 2 GB of GDDR5 on a 256-bit bus, delivering 192.3 GB/s of bandwidth. The GTX 1660 SUPER has 6 GB of GDDR6 on a 192-bit bus, delivering 336.0 GB/s. Despite the narrower bus, the newer memory type and higher effective speed (14 Gbps versus 6 Gbps) more than compensate. The GTX 680's memory clocks at 1502 MHz (6 Gbps effective), while the GTX 1660 SUPER runs at 1750 MHz (14 Gbps effective).

Compute resources differ in configuration. The GTX 680 has 1,536 shading units, 128 texture mapping units, and 32 ROPs. The GTX 1660 SUPER has 1,408 shading units, 88 TMUs, and 48 ROPs. Fewer shaders and TMUs, but more ROPs, and the clock speeds are much higher: the GTX 680's base is 1006 MHz with a boost of 1058 MHz, while the GTX 1660 SUPER starts at 1530 MHz and boosts to 1785 MHz. The result is that the GTX 1660 SUPER achieves 5.027 TFLOPS FP32 versus the GTX 680's 3.250 TFLOPS, and its pixel rate is 85.68 GPixel/s versus 33.86 GPixel/s. The texture rate is closer: 157.1 GTexel/s versus 135.4 GTexel/s.

The GTX 1660 SUPER also supports FP16 compute at 10.05 TFLOPS (2:1 ratio), which the GTX 680 lacks entirely. API support differs as well: the GTX 680 lists DirectX 12 (11_0) and Vulkan 1.2.175, while the GTX 1660 SUPER lists DirectX 12 (12_1) and Vulkan 1.4. Both support OpenGL 4.6. Neither card has RT cores or tensor cores.

Power and physical characteristics follow the generational trend. The GTX 680 has a TDP of 195 W with a suggested PSU of 450 W and two 6-pin power connectors. The GTX 1660 SUPER draws 125 W with a suggested PSU of 300 W and a single 8-pin connector. The GTX 680 is longer at 256 mm (10.1 inches) versus 229 mm (9 inches) for the GTX 1660 SUPER. Both are dual-slot, and both are 111 mm tall. The GTX 680 is 38 mm wide versus 35 mm for the GTX 1660 SUPER.

FAQ

Q: Which card is faster in compute workloads?

A: The GTX 1660 SUPER is significantly faster. In Geekbench OpenCL it scores 52,490 versus 16,906 for the GTX 680, a 67.8 percent advantage. In Geekbench Vulkan, the gap is 69.1 percent in favor of the GTX 1660 SUPER (57,102 versus 17,646).

Q: Do these cards support modern graphics APIs?

A: Both support DirectX 12 and Vulkan, but at different feature levels. The GTX 680 is limited to DirectX 12 (11_0) and Vulkan 1.2.175, while the GTX 1660 SUPER supports DirectX 12 (12_1) and Vulkan 1.4. Both run OpenGL 4.6.

Q: What are the memory specifications of each card?

A: The GTX 680 has 2 GB of GDDR5 on a 256-bit bus with 192.3 GB/s bandwidth. The GTX 1660 SUPER has 6 GB of GDDR6 on a 192-bit bus with 336.0 GB/s bandwidth. The effective memory speed is 6 Gbps for the GTX 680 and 14 Gbps for the GTX 1660 SUPER.

Q: How do their power requirements compare?

A: The GTX 680 has a TDP of 195 W and requires a 450 W PSU with two 6-pin connectors. The GTX 1660 SUPER has a TDP of 125 W, requires a 300 W PSU, and uses one 8-pin connector.

Q: Which card has more shading units?

A: The GTX 680 has 1,536 shading units, while the GTX 1660 SUPER has 1,408. However, the GTX 1660 SUPER runs at much higher clocks, so it delivers more overall compute performance.

Q: Are either of these cards still in production?

A: No. Both are listed as end-of-life in the database. The GTX 680 launched on 2012-03-21, and the GTX 1660 SUPER launched on 2019-10-28.

Specification Differences

The following fields differ between the two cards:

• Architecture: Kepler (GK104) versus Turing (TU116)
• Process node: 28 nm versus 12 nm
• Transistors: 3,540 million versus 6,600 million
• Die size: 294 mm² versus 284 mm²
• Transistor density: 12.0M / mm² versus 23.2M / mm²
• Base clock: 1006 MHz versus 1530 MHz
• Boost clock: 1058 MHz versus 1785 MHz
• Memory clock: 1502 MHz (6 Gbps effective) versus 1750 MHz (14 Gbps effective)
• Memory size: 2 GB versus 6 GB
• Memory type: GDDR5 versus GDDR6
• Memory bus width: 256 bit versus 192 bit
• Memory bandwidth: 192.3 GB/s versus 336.0 GB/s
• Shading units: 1536 versus 1408
• TMUs: 128 versus 88
• ROPs: 32 versus 48
• Pixel rate: 33.86 GPixel/s versus 85.68 GPixel/s
• Texture rate: 135.4 GTexel/s versus 157.1 GTexel/s
• FP32 performance: 3.250 TFLOPS versus 5.027 TFLOPS
• FP16 performance: Not supported versus 10.05 TFLOPS (2:1)
• TDP: 195 W versus 125 W
• Power connectors: 2x 6-pin versus 1x 8-pin
• Suggested PSU: 450 W versus 300 W
• Display outputs: 2x DVI, 1x HDMI 1.4a, 1x DisplayPort 1.2 versus 1x DVI, 1x HDMI 2.0, 1x DisplayPort 1.4a
• DirectX support: 12 (11_0) versus 12 (12_1)
• Vulkan support: 1.2.175 versus 1.4
• Dimensions: 256 mm x 111 mm x 38 mm versus 229 mm x 111 mm x 35 mm
• Release date: 2012-03-21 versus 2019-10-28
• Launch MSRP: 499 USD versus 229 USD

Where Each One Wins

The GTX 680 wins in zero direct head-to-head tests. Its only recorded strengths are in the database's overall percentile ranking, where it sits at 55 versus the GTX 1660 SUPER's 53, and in its average benchmark score of 14,150 versus 12,986. These are artifacts of test selection rather than genuine performance advantages. The GTX 680 also has more shading units (1,536 versus 1,408), more TMUs (128 versus 88), and a wider memory bus (256 bit versus 192 bit), but these do not translate into wins in any shared benchmark.

The GTX 1660 SUPER wins both shared benchmarks decisively. Its advantages include higher clocks, more memory capacity, faster memory technology, higher bandwidth, more ROPs, higher pixel and texture rates, higher FP32 throughput, FP16 support, newer API versions, lower power draw, and a smaller physical footprint. The GTX 1660 SUPER also has a lower suggested PSU requirement (300 W versus 450 W), which makes it easier to integrate into existing systems.

For gaming and general compute, the GTX 1660 SUPER is the clear winner. The 6 GB memory capacity is three times larger, which matters for modern textures and larger datasets. The 336.0 GB/s bandwidth versus 192.3 GB/s gives it a major edge in memory-bound scenarios. The higher pixel rate (85.68 versus 33.86 GPixel/s) indicates faster rasterization. The newer DirectX 12 (12_1) and Vulkan 1.4 support ensure better compatibility with current software.

The GTX 680's only conceivable niche is legacy system compatibility. It uses older display outputs (HDMI 1.4a, DisplayPort 1.2) that may be required for very old monitors, but the GTX 1660 SUPER still offers DVI, HDMI 2.0, and DisplayPort 1.4a, so this is a minor distinction. The GTX 680's higher TDP of 195 W might be relevant for someone specifically seeking a card that draws more power, but that is not a practical advantage.

The Verdict

The data is unambiguous. The NVIDIA GeForce GTX 1660 SUPER outperforms the GTX 680 by 67.8 percent in OpenCL and 69.1 percent in Vulkan. It delivers more than double the FP32 compute (5.027 versus 3.250 TFLOPS), more than double the pixel rate (85.68 versus 33.86 GPixel/s), and nearly double the memory bandwidth (336.0 versus 192.3 GB/s). It has triple the memory capacity (6 GB versus 2 GB), newer memory technology (GDDR6 versus GDDR5), and superior API support (DirectX 12_1 and Vulkan 1.4 versus DirectX 11_0 and Vulkan 1.2.175). It does all of this while drawing 70 W less power (125 W versus 195 W) and requiring a smaller PSU (300 W versus 450 W).

The GTX 680's launch MSRP was 499 USD, while the GTX 1660 SUPER launched at 229 USD. Even without considering price-to-performance ratios, the GTX 1660 SUPER is the superior product in every measured category. The only database metric where the GTX 680 appears ahead is the aggregate average score and percentile, which are skewed by the differing benchmark suites. Any builder choosing between these two cards for a new or upgraded system should select the GTX 1660 SUPER. The GTX 680 should only be considered if the system absolutely requires Kepler-era display outputs or if the card is already installed and the workload does not benefit from the GTX 1660 SUPER's advantages. For all practical purposes, the GTX 1660 SUPER is the definitive choice.