NVIDIA GeForce GTX 1650 SUPER vs NVIDIA GeForce GTX 670 Comparison

Where Each One Wins

The benchmark data splits cleanly between these two NVIDIA cards, and the wins are one-sided. The NVIDIA GeForce GTX 1650 SUPER wins both recorded head-to-head tests outright, while the GTX 670 does not win a single comparison in the database. This is not a close contest in raw compute workloads; the newer Turing-based card dominates in every measurable category where both were tested.

In Geekbench OpenCL, the GTX 1650 SUPER scores 43,875 against the GTX 670's 15,341. That is a 65% deficit for the older card, meaning the GTX 1650 SUPER delivers roughly 2.9 times the compute throughput in this API. The gap is even larger in Geekbench Vulkan: the GTX 1650 SUPER posts 50,519 versus 15,553 for the GTX 670, a 69.2% difference. For any application that leverages Vulkan or OpenCL, the GTX 1650 SUPER is the clear choice.

The GTX 670's strengths, if any, lie outside the direct head-to-head tests. Its average benchmark score across all recorded tests is 12,773, which places it at the 52nd percentile among all GPUs in the database. The GTX 1650 SUPER averages 11,047, sitting at the 50th percentile. So while the GTX 670 loses the two direct comparisons, its overall average is actually 15.6% higher than the GTX 1650 SUPER's average. This is because the GTX 670's benchmark set includes different tests (Geekbench Metal, OpenCL, and Vulkan), while the GTX 1650 SUPER's set includes a broader range of Passmark tests, many of which are legacy DirectX workloads where it scores modestly.

The GTX 670 also shows competitive positioning against its nearest rivals. Its average score of 12,773 is within 0.4% of the GTX 590 (12,830) and the Radeon Pro 455 (12,831), and within 0.6% of the FirePro W5100 (12,847). It actually edges out the RX 7600M XT (12,710) by 0.5%. The GTX 1650 SUPER, by contrast, sits just 0.2% below the Radeon RX 550 (11,075) and 1.5% above the GeForce MX350 (10,883). In terms of raw average scores, the GTX 670 is the more balanced performer across its test suite, but the GTX 1650 SUPER is decisively faster in the modern compute APIs that matter for current workloads.

The Verdict

The data points to one conclusion: choose the GTX 1650 SUPER for any modern compute or graphics workload that uses Vulkan or OpenCL. Its scores are more than double the GTX 670's in both head-to-head tests, and it does so while consuming far less power. The GTX 1650 SUPER has a TDP of 100 W versus 170 W for the GTX 670, and it requires only a single 6-pin power connector with a suggested 300 W PSU, compared to the GTX 670's dual 6-pin connectors and 450 W PSU recommendation.

The GTX 670 does have a claim for legacy scenarios. Its average benchmark score of 12,773 beats the GTX 1650 SUPER's 11,047, and it holds a higher percentile ranking (52nd versus 50th). If the workload is exclusively older DirectX titles or applications that do not use Vulkan or OpenCL, the GTX 670's higher average suggests it may hold up better in those specific environments. But the head-to-head data shows the GTX 1650 SUPER winning both recorded tests by massive margins, and those are the tests the database uses for direct comparison.

For gamers or professionals running modern software, the GTX 1650 SUPER is the only rational pick. It offers 4 GB of GDDR6 memory versus 2 GB of GDDR5, a 12 nm Turing architecture versus 28 nm Kepler, and supports DirectX 12 (12_1) versus the GTX 670's DirectX 12 (11_0). The GTX 670 is end-of-life hardware from 2012, while the GTX 1650 SUPER, though also end-of-life, launched in 2019 with substantially newer feature support. The verdict is unambiguous: the GTX 1650 SUPER wins on performance, efficiency, and modern API support.

Head-to-Head Benchmarks

The database records two direct comparisons, and the GTX 1650 SUPER wins both by overwhelming margins.

In Geekbench OpenCL, the GTX 1650 SUPER scores 43,875 against the GTX 670's 15,341. The delta is 65% in favor of the GTX 1650 SUPER. This is a compute-focused test that stresses raw throughput across the GPU's shading units. The GTX 1650 SUPER's FP32 performance is listed at 4.416 TFLOPS, while the GTX 670 manages 2.634 TFLOPS. That 67.6% raw compute advantage translates almost directly into the OpenCL score gap. The GTX 1650 SUPER also has a higher texture rate (138.0 GTexel/s versus 109.8 GTexel/s) and nearly double the pixel rate (55.20 GPixel/s versus 27.44 GPixel/s).

In Geekbench Vulkan, the gap widens further. The GTX 1650 SUPER scores 50,519 versus 15,553 for the GTX 670, a 69.2% difference. Vulkan is a low-level API that benefits from modern architecture design, and the Turing-based GTX 1650 SUPER has clear advantages. It supports Vulkan 1.4, while the GTX 670 is limited to Vulkan 1.2.175. The GTX 1650 SUPER's higher boost clock (1725 MHz versus 980 MHz) and its 12 Gbps effective memory speed (versus 6 Gbps on the GTX 670) contribute to this result.

The GTX 670's only statistical consolation is its average benchmark score of 12,773, which is higher than the GTX 1650 SUPER's 11,047. But that average includes tests that favor the older card, such as Geekbench Metal (7,424 for the GTX 670, a test the GTX 1650 SUPER does not have recorded). When both cards are measured on the same workloads, the GTX 1650 SUPER wins decisively in both cases. The data shows no scenario where the GTX 670 beats the GTX 1650 SUPER in a direct head-to-head test.

FAQ

Q: Which card has the higher average benchmark score?

A: The GTX 670 has an average benchmark score of 12,773, while the GTX 1650 SUPER averages 11,047. The GTX 670's average is 15.6% higher, despite losing both head-to-head tests.

Q: How much faster is the GTX 1650 SUPER in Vulkan?

A: The GTX 1650 SUPER scores 50,519 in Geekbench Vulkan, while the GTX 670 scores 15,553. That is a 69.2% advantage for the GTX 1650 SUPER.

Q: What memory configurations do the two cards use?

A: The GTX 670 has 2 GB of GDDR5 memory on a 256-bit bus with 192.3 GB/s bandwidth. The GTX 1650 SUPER has 4 GB of GDDR6 memory on a 128-bit bus with 192.0 GB/s bandwidth.

Q: Which card is more power efficient?

A: The GTX 1650 SUPER has a TDP of 100 W and requires a single 6-pin connector with a 300 W suggested PSU. The GTX 670 has a TDP of 170 W, dual 6-pin connectors, and a 450 W suggested PSU.

Q: Does the GTX 670 win any benchmark against the GTX 1650 SUPER?

A: No. In the two recorded head-to-head tests (Geekbench OpenCL and Geekbench Vulkan), the GTX 1650 SUPER wins both. The GTX 670 has zero wins in the head-to-head comparison.

Q: What are the nearest rivals for each card?

A: The GTX 670's nearest rivals include the GTX 590 (0.4% lower average score), Radeon Pro 455 (0.4% lower), RX 7600M XT (0.5% higher), and FirePro W5100 (0.6% lower). The GTX 1650 SUPER's nearest rivals include the Radeon RX 550 (0.2% higher), RTX PRO 6000D Blackwell Max-Q (0.4% higher), RTX PRO 6000 Blackwell Max-Q (0.4% higher), and GeForce MX350 (1.5% lower).

Architecture Differences

The two cards come from entirely different architectural generations. The GTX 670 uses the GK104 chip based on Kepler architecture, built on a 28 nm process at TSMC. It packs 3,540 million transistors into a 294 mm² die, yielding a transistor density of 12.0 million per mm². The GTX 1650 SUPER uses the TU116 chip based on Turing architecture, built on a 12 nm process, also at TSMC. It contains 6,600 million transistors on a 284 mm² die, for a density of 23.2 million per mm². The GTX 1650 SUPER has nearly double the transistor count (86.4% more) in a slightly smaller die.

The shading unit counts are close: the GTX 670 has 1,344 shading units, while the GTX 1650 SUPER has 1,280. However, the GTX 670 has more texture mapping units (112 versus 80) and the same number of raster operations units (32 on both). The GTX 1650 SUPER compensates with much higher clocks: 1,530 MHz base and 1,725 MHz boost, versus 915 MHz base and 980 MHz boost on the GTX 670. This clock advantage drives its FP32 output to 4.416 TFLOPS, compared to 2.634 TFLOPS on the GTX 670.

The GTX 1650 SUPER also supports FP16 compute at 8.832 TFLOPS (2:1 rate), a feature the GTX 670 does not list. Neither card has ray tracing or tensor cores. The GTX 670 supports DirectX 12 (11_0), while the GTX 1650 SUPER supports DirectX 12 (12_1). Vulkan support also differs: 1.2.175 on the GTX 670 versus 1.4 on the GTX 1650 SUPER. OpenGL support is identical at 4.6.

Specification Differences

The memory subsystem is a key differentiator. The GTX 670 offers 2 GB of GDDR5 on a 256-bit bus, delivering 192.3 GB/s bandwidth. The GTX 1650 SUPER offers 4 GB of GDDR6 on a 128-bit bus, delivering 192.0 GB/s. The effective memory speed is 6 Gbps on the GTX 670 versus 12 Gbps on the GTX 1650 SUPER, but the wider bus on the GTX 670 equalizes bandwidth almost exactly.

The power requirements diverge sharply. The GTX 670 consumes 170 W with dual 6-pin power connectors and a suggested 450 W PSU. The GTX 1650 SUPER consumes 100 W with a single 6-pin connector and a suggested 300 W PSU. Both are dual-slot cards, but the GTX 670 is slightly longer at 241 mm (9.5 inches) versus 229 mm (9 inches). Both are 111 mm (4.4 inches) tall, and the GTX 670 is 38 mm (1.5 inches) wide versus 35 mm (1.4 inches) for the GTX 1650 SUPER.

Display outputs differ: the GTX 670 has 2x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2. The GTX 1650 SUPER has 1x DVI, 1x HDMI 2.0, and 1x DisplayPort 1.4a. Both use PCIe 3.0 x16. The GTX 670 launched on May 9, 2012 with a launch MSRP of 399 USD, while the GTX 1650 SUPER launched on November 21, 2019 with a launch MSRP of 159 USD. Both are marked end-of-life in the database.