AMD Radeon RX 7900 XT vs NVIDIA GeForce GTX 1660 Comparison

FAQ

Q: Which GPU wins the majority of head-to-head benchmark comparisons?

A: The AMD Radeon RX 7900 XT wins 9 out of 10 recorded head-to-head benchmarks against the NVIDIA GeForce GTX 1660. The only exception is Geekbench OpenCL, where the GTX 1660 posts a higher score.

Q: How large is the performance gap in the 3DMark Steel Nomad DX12 test?

A: The RX 7900 XT scores 5623 points versus 1065 points for the GTX 1660, a delta of 428%. This is the largest single-test advantage recorded between the two cards.

Q: Does the GTX 1660 win any compute-oriented benchmark?

A: Yes, in Geekbench OpenCL the GTX 1660 scores 47850 against the RX 7900 XT's 12756, a 73.3% advantage for NVIDIA. This result is anomalous compared to the PassMark GPU compute test, where AMD leads by 239.1%.

Q: How do the two cards compare in overall average benchmark score?

A: The RX 7900 XT has an average benchmark score of 13745, while the GTX 1660 averages 11680. The AMD card sits at the 55th percentile of all GPUs, and the NVIDIA card at the 51st percentile.

Q: What are the nearest rivals for each card according to average score?

A: For the RX 7900 XT, the nearest rival is the AMD Radeon 660M at 13812 (0.5% higher), followed by the NVIDIA RTX A2000 Mobile at 13821 (0.5% higher). For the GTX 1660, the AMD Radeon RX 7800 XT at 11627 is 0.5% lower, and the AMD Radeon RX 6500 XT at 11842 is 1.4% higher.

Q: Which card has the higher launch MSRP?

A: The AMD Radeon RX 7900 XT has a launch MSRP of 899 USD, while the NVIDIA GeForce GTX 1660 has a launch MSRP of 219 USD.

Architecture Differences

The AMD Radeon RX 7900 XT is built on the RDNA 3.0 architecture using the Navi 31 chip, codenamed Plum Bonito, and belongs to the Navi III (RX 7000) generation. It is fabricated on a 5 nm process at TSMC, with 57,700 million transistors on a 529 mm² die, yielding a transistor density of 109.1M per mm². The NVIDIA GeForce GTX 1660 uses the Turing architecture on the TU116 chip, part of the GeForce 16 generation. It is built on a 12 nm process at TSMC, with 6,600 million transistors on a 284 mm² die, for a density of 23.2M per mm². The process and density differences are substantial: the AMD chip packs over eight times the transistor count on a die that is less than twice the area.

The RX 7900 XT features 5,376 shading units, 336 texture mapping units, and 192 raster operation pipelines. It also includes 84 ray tracing cores, a feature class absent from the GTX 1660, which has no ray tracing hardware and no tensor cores. The GTX 1660 offers 1,408 shading units, 88 TMUs, and 48 ROPs. The AMD card's raw throughput numbers reflect this scale: 51.48 TFLOPS FP32 versus 5.027 TFLOPS for NVIDIA, 804.4 GTexel/s versus 157.1 GTexel/s, and 459.6 GPixel/s versus 85.68 GPixel/s. FP16 performance is 103.0 TFLOPS (2:1) for AMD and 10.05 TFLOPS (2:1) for NVIDIA, a near 10-fold gap.

Memory architecture diverges sharply. The RX 7900 XT uses 20 GB of GDDR6 on a 320-bit bus, delivering 800.0 GB/s of bandwidth, with memory clocked at 2500 MHz (20 Gbps effective). The GTX 1660 has 6 GB of GDDR5 on a 192-bit bus, providing 192.1 GB/s, with memory at 2001 MHz (8 Gbps effective). The AMD card's memory bandwidth is more than four times higher, and its capacity is more than three times larger. Clock behavior also differs: the RX 7900 XT has a base clock of 1387 MHz, a boost of 2394 MHz, and a game clock of 2025 MHz, while the GTX 1660 runs at 1530 MHz base and 1785 MHz boost, with no game clock listed. Despite the lower base clock, the AMD part boosts substantially higher.

Interface and connectivity reflect their respective generations. The RX 7900 XT uses PCIe 4.0 x16, while the GTX 1660 uses PCIe 3.0 x16. Display outputs differ as well: the AMD card provides 1x HDMI 2.1a, 2x DisplayPort 2.1, and 1x USB Type-C, whereas the GTX 1660 offers 1x DVI, 1x HDMI 2.0, and 1x DisplayPort 1.4a. API support is another differentiator: the RX 7900 XT supports DirectX 12 Ultimate (12_2), while the GTX 1660 only reaches DirectX 12 (12_1). Both support OpenGL 4.6 and Vulkan 1.4. Power draw and power delivery scale accordingly: the RX 7900 XT is rated at 300 W with a suggested PSU of 700 W and dual 8-pin connectors, versus 120 W, 300 W suggested PSU, and a single 8-pin connector for the GTX 1660. Both are dual-slot cards, but the RX 7900 XT is longer at 276 mm versus 229 mm, slightly shorter in height at 110 mm versus 111 mm, and thicker at 51 mm versus 35 mm.

Head-to-Head Benchmarks

The RX 7900 XT dominates the head-to-head results, winning 9 of 10 tests. The largest margin is in 3DMark Steel Nomad DX12, where the AMD card scores 5623 versus 1065, a 428% lead. This test highlights the generational gulf in modern DX12 rasterization and compute workloads. In PassMark GPU compute, the RX 7900 XT scores 16828 against 4963, a 239.1% advantage, confirming that the AMD card's massive FP32 and FP16 throughput translates directly into compute-heavy tasks. PassMark DirectX 11 shows a 330.4% lead, with scores of 340 versus 79, and PassMark DirectX 10 shows a 159% lead, 158 versus 61. The pattern across the PassMark suite is consistent: the RX 7900 XT leads by 81.9% in DirectX 9 (322 versus 177), 124.5% in DirectX 12 (110 versus 49), and 149.1% in the overall G3D test (29009 versus 11646). In PassMark G2D, the AMD card leads 1232 to 776, a 58.8% margin, indicating a strong advantage even in 2D graphics workloads. In Geekbench Vulkan, the RX 7900 XT scores 71068 versus 50137, a 41.7% lead.

The single test where the GTX 1660 wins is Geekbench OpenCL, where it scores 47850 against the RX 7900 XT's 12756, a 73.3% advantage. This result is striking given the PassMark GPU compute result going the other way by a wide margin. The discrepancy suggests that the OpenCL workload in Geekbench responds differently to the two architectures, possibly favoring the NVIDIA driver stack or the Turing compute pipeline in that specific test. However, it is an outlier: every other recorded benchmark, including Vulkan and DirectX variants, lands in favor of the AMD part. The average benchmark score difference further contextualizes the matchup, with the RX 7900 XT at 13745 and the GTX 1660 at 11680. The RX 7900 XT's nearest rivals in the database, the AMD Radeon 660M and NVIDIA RTX A2000 Mobile, sit within 0.5% of its average score, while the GTX 1660's nearest rival cluster includes the RX 7800 XT at 0.5% lower and the RX 6500 XT at 1.4% higher.

Specification Differences

The two cards differ in nearly every measured specification category. Process node: 5 nm for AMD versus 12 nm for NVIDIA. Transistor count: 57,700 million versus 6,600 million. Die size: 529 mm² versus 284 mm². Transistor density: 109.1M per mm² versus 23.2M per mm². Shading units: 5,376 versus 1,408. TMUs: 336 versus 88. ROPs: 192 versus 48. Ray tracing cores: 84 on the AMD card, none on the NVIDIA card. Memory size: 20 GB versus 6 GB. Memory type: GDDR6 versus GDDR5. Bus width: 320 bit versus 192 bit. Bandwidth: 800.0 GB/s versus 192.1 GB/s. Pixel rate: 459.6 GPixel/s versus 85.68 GPixel/s. Texture rate: 804.4 GTexel/s versus 157.1 GTexel/s. FP32: 51.48 TFLOPS versus 5.027 TFLOPS. FP16: 103.0 TFLOPS versus 10.05 TFLOPS. TDP: 300 W versus 120 W. Power connectors: 2x 8-pin versus 1x 8-pin. Suggested PSU: 700 W versus 300 W. Bus interface: PCIe 4.0 x16 versus PCIe 3.0 x16. Display outputs: 1x HDMI 2.1a, 2x DisplayPort 2.1, 1x USB Type-C versus 1x DVI, 1x HDMI 2.0, 1x DisplayPort 1.4a. DirectX support: 12 Ultimate (12_2) versus 12 (12_1). Dimensions: 276 mm x 110 mm x 51 mm versus 229 mm x 111 mm x 35 mm. Release dates differ by several years, with the RX 7900 XT released after the GTX 1660, and each card is end-of-life in production status. The RX 7900 XT has a launch MSRP of 899 USD, while the GTX 1660 has a launch MSRP of 219 USD.

The Verdict

The benchmark data is unambiguous for almost the entire test suite. The AMD Radeon RX 7900 XT is the faster card in 9 of 10 recorded tests, with leads ranging from 41.7% in Geekbench Vulkan to 428% in 3DMark Steel Nomad DX12. Its average benchmark score of 13745 is higher than the GTX 1660's 11680, and it sits at the 55th percentile of all GPUs versus the 51st percentile for NVIDIA. The RX 7900 XT is the choice for anyone prioritizing modern DX12, Vulkan, or compute-heavy workloads, as evidenced by its dominant PassMark GPU compute and 3DMark results. It also brings architectural features the GTX 1660 lacks entirely, including ray tracing cores, DirectX 12 Ultimate support, and a substantially larger memory subsystem with 20 GB of GDDR6 on a 320-bit bus.

The GTX 1660's single victory in Geekbench OpenCL, at 73.3% ahead, is notable but isolated. It does not carry over to any other compute or graphics test in the database. The NVIDIA card's advantage in that one workload, along with its lower power draw of 120 W and smaller physical footprint, may matter for specific use cases, but the data does not support a broader performance claim. The GTX 1660's nearest rivals in the database, including the RX 7800 XT and RX 6500 XT, all sit within a narrow band around its average score, suggesting it competes in a much lower performance tier.

For a user selecting between these two specific cards based strictly on recorded measurements, the RX 7900 XT is the stronger performer in the overwhelming majority of tests, with its largest wins coming in exactly the areas where modern workloads are moving: DX12, Vulkan, and GPGPU compute. The GTX 1660 remains competitive only in a single OpenCL benchmark and offers lower power requirements and a smaller physical size. The database records no scenario outside that OpenCL test where the GTX 1660 leads, so the verdict follows the data: the RX 7900 XT for general and compute performance, the GTX 1660 only for the specific OpenCL workload where it holds the edge.