AMD Radeon RX 5500 XT vs NVIDIA GeForce GTX 670 Comparison

Where Each One Wins

The benchmark data presents an unusually one-sided picture. Across the three directly comparable tests, the AMD Radeon RX 5500 XT claims victory in every single instance. The NVIDIA GeForce GTX 670 records zero wins in the head-to-head comparison. This is not a case of subtle margins or trade-offs; the RX 5500 XT dominates the GTX 670 by margins ranging from roughly 184% to over 638%.

The most dramatic separation occurs in the Geekbench Metal test. The RX 5500 XT scores 54,842 points against the GTX 670's 7,424 points, a delta of 638.7%. This is the kind of gap that suggests fundamental architectural differences rather than mere clock speed advantages. Metal is Apple's graphics API, and the RX 5500 XT's RDNA 1.0 architecture appears far better optimized for it, or simply has enough raw compute headroom to run away with the test.

In OpenCL, the RX 5500 XT posts 46,965 points versus 15,341 points for the GTX 670, a 206.1% advantage. OpenCL is a cross-platform compute standard, so this result reflects general-purpose GPU compute capability. The RX 5500 XT's FP32 throughput of 5.196 TFLOPS compared to the GTX 670's 2.634 TFLOPS explains much of this gap; the AMD card has nearly twice the raw floating-point throughput.

Vulkan shows the narrowest margin of the three, but it is still decisive. The RX 5500 XT scores 44,191 against 15,553, a 184.1% lead. Vulkan is a low-overhead graphics API, and the RX 5500 XT's support for Vulkan 1.4 versus the GTX 670's Vulkan 1.2.175 suggests the newer architecture handles modern API workloads substantially better.

The GTX 670 has no benchmark category where it wins. Its only competitive angle is the broader set of legacy DirectX tests in the Passmark suite, but those results exist only for the RX 5500 XT in the database, so no direct comparison can be drawn. The GTX 670's pixel rate of 27.44 GPixel/s and texture rate of 109.8 GTexel/s are both far below the RX 5500 XT's 59.04 GPixel/s and 162.4 GTexel/s, which aligns with the observed benchmark outcomes.

The Verdict

The data is unambiguous about who should pick which card. The AMD Radeon RX 5500 XT is the superior choice in every measurable category available in the database. It wins all three head-to-head benchmark tests, holds a 57th percentile ranking among all GPUs compared to the GTX 670's 52nd percentile, and posts an average benchmark score of 14,692 against the GTX 670's 12,773.

The RX 5500 XT sits in interesting company among its nearest rivals. It trails the NVIDIA GeForce GTX 770 by a mere 0.4%, leads the NVIDIA Quadro M2000 by 1.1%, and edges out the AMD Radeon RX Vega 11 by 2.1%. This places it in a performance tier that is competitive with older high-end cards while using a fraction of the power envelope.

The GTX 670, meanwhile, finds itself grouped with the NVIDIA GeForce GTX 590 (0.4% behind), the AMD Radeon Pro 455 (0.4% behind), and the AMD FirePro W5100 (0.6% behind). Its 12773 average score is respectable for its era, but it simply cannot keep pace with the RX 5500 XT's modern architecture.

For anyone choosing between these two specific cards, the RX 5500 XT is the only rational pick based on recorded data. It wins on compute, graphics API support, memory bandwidth, and overall average score. The GTX 670's only advantage is historical: it launched earlier and at a higher MSRP, but the database records no performance category where it comes out ahead.

Head-to-Head Benchmarks

The Geekbench Metal test delivers the single largest margin in the comparison. The RX 5500 XT's 54,842 score dwarfs the GTX 670's 7,424, producing a 638.7% delta. This is not a close contest by any measure. The RX 5500 XT's RDNA 1.0 architecture, built on a 7 nm process, clearly offers Metal compute capabilities that the 28 nm Kepler architecture cannot approach.

Geekbench OpenCL narrows the gap but remains heavily lopsided. The RX 5500 XT scores 46,965 versus 15,341 for the GTX 670, a 206.1% difference. The GTX 670's FP32 throughput of 2.634 TFLOPS is roughly half the RX 5500 XT's 5.196 TFLOPS, and the benchmark result reflects that ratio closely. The RX 5500 XT also offers FP16 performance of 10.39 TFLOPS (2:1), a feature the GTX 670 lacks entirely, which may contribute to compute-heavy workloads.

Geekbench Vulkan delivers the closest relative contest, but the outcome is still decisive. The RX 5500 XT's 44,191 score beats the GTX 670's 15,553 by 184.1%. Vulkan performance tends to reward modern architectures with better driver optimization and hardware feature support. The RX 5500 XT supports Vulkan 1.4, while the GTX 670 is limited to Vulkan 1.2.175, which likely explains part of the gap.

Beyond the head-to-head tests, the broader benchmark set reinforces the RX 5500 XT's advantage. Its Passmark G3D score of 9,083 and GPU compute score of 4,446 indicate solid general performance. The GTX 670 has no recorded Passmark scores in the database, so these cannot be compared directly.

The average benchmark scores tell the same story. The RX 5500 XT averages 14,692 across all recorded tests, while the GTX 670 averages 12,773. The RX 5500 XT sits at the 57th percentile of all GPUs, the GTX 670 at the 52nd. Every metric in the database points in the same direction.

FAQ

Q: Which card is faster in Geekbench Metal?

A: The AMD Radeon RX 5500 XT scores 54,842 compared to the NVIDIA GeForce GTX 670's 7,424, a 638.7% advantage.

Q: Does the GTX 670 win any benchmark against the RX 5500 XT?

A: No. The RX 5500 XT wins all three head-to-head tests: Geekbench Metal, OpenCL, and Vulkan. The GTX 670 has zero recorded wins.

Q: How do the average benchmark scores compare?

A: The RX 5500 XT has an average benchmark score of 14,692, while the GTX 670 averages 12,773. The RX 5500 XT ranks at the 57th percentile of all GPUs, the GTX 670 at the 52nd.

Q: Which card has better Vulkan support?

A: The RX 5500 XT supports Vulkan 1.4 and scores 44,191 in the Geekbench Vulkan test. The GTX 670 supports Vulkan 1.2.175 and scores 15,553, a 184.1% difference.

Q: How does the RX 5500 XT compare to its nearest rivals?

A: The RX 5500 XT trails the NVIDIA GeForce GTX 770 by 0.4%, leads the NVIDIA Quadro M2000 by 1.1%, and leads the AMD Radeon RX Vega 11 by 2.1%.

Q: What is the GTX 670's closest rival in the database?

A: The GTX 670 is 0.4% behind the NVIDIA GeForce GTX 590 and the AMD Radeon Pro 455, and 0.6% behind the AMD FirePro W5100. It leads the AMD Radeon RX 7600M XT by 0.5%.

Architecture Differences

The two cards represent fundamentally different eras of GPU design. The RX 5500 XT uses the Navi 14 chip built on RDNA 1.0 architecture, while the GTX 670 uses the GK104 chip built on Kepler architecture. The manufacturing process alone tells the story: the RX 5500 XT is fabricated on a 7 nm process at TSMC, while the GTX 670 uses a 28 nm process, also at TSMC.

The transistor counts reflect this generational leap. The RX 5500 XT packs 6,400 million transistors into a die size of 158 mm², yielding a transistor density of 40.5 million per square millimeter. The GTX 670 contains 3,540 million transistors across a much larger 294 mm² die, for a density of just 12.0 million per square millimeter. The RX 5500 XT achieves more than three times the transistor density of the GTX 670.

Memory architecture differs significantly. The RX 5500 XT uses 4 GB of GDDR6 memory on a 128-bit bus, delivering 224.0 GB/s of bandwidth. The GTX 670 uses 2 GB of GDDR5 memory on a 256-bit bus, delivering 192.3 GB/s. The RX 5500 XT achieves higher bandwidth despite a narrower bus, thanks to the newer memory technology.

Compute resources show a mixed picture. The RX 5500 XT has 1,408 shading units and 88 texture mapping units. The GTX 670 has 1,344 shading units but 112 TMUs. Both have 32 ROPs. The RX 5500 XT's higher shading unit count, combined with its much higher clocks, produces substantially better throughput figures.

Clock speeds heavily favor the RX 5500 XT. It runs at a base clock of 1607 MHz, a game clock of 1717 MHz, and a boost clock of 1845 MHz. The GTX 670 operates at 915 MHz base and 980 MHz boost. The RX 5500 XT's memory runs at 1750 MHz (14 Gbps effective), while the GTX 670's memory runs at 1502 MHz (6 Gbps effective).

The RX 5500 XT supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. The GTX 670 supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. The DirectX feature level difference (12_1 versus 11_0) indicates the RX 5500 XT supports more modern rendering features.

Specification Differences

The RX 5500 XT draws 130 W TDP with a single 8-pin power connector and a suggested 300 W PSU. The GTX 670 draws 170 W TDP with two 6-pin connectors and a suggested 450 W PSU. The RX 5500 XT delivers substantially more performance while consuming less power.

The RX 5500 XT uses PCIe 4.0 x8, while the GTX 670 uses PCIe 3.0 x16. The newer PCIe standard offers higher bandwidth per lane, which partially compensates for the RX 5500 XT's fewer lanes.

Physical dimensions differ notably. The RX 5500 XT measures 180 mm (7.1 inches) in length. The GTX 670 is longer at 241 mm (9.5 inches), with additional dimensions recorded: 111 mm (4.4 inches) in height and 38 mm (1.5 inches) in width. Both are dual-slot cards.

Display outputs reflect their respective eras. The RX 5500 XT offers one HDMI 2.0b port and three DisplayPort 1.4a ports. The GTX 670 provides two DVI ports, one HDMI 1.4a port, and one DisplayPort 1.2 port.

The RX 5500 XT has a launch MSRP of 169 USD. The GTX 670 has a launch MSRP of 399 USD. The RX 5500 XT released on 2019-12-11, while the GTX 670 released on 2012-05-09. Both are end-of-life products. The RX 5500 XT's predecessor is Vega and its successor is Navi II. The GTX 670's predecessor is GeForce 500 and its successor is GeForce 700.

The RX 5500 XT's FP32 performance of 5.196 TFLOPS and FP16 performance of 10.39 TFLOPS (2:1) dwarf the GTX 670's 2.634 TFLOPS FP32 with no FP16 support recorded. Pixel rate favors the RX 5500 XT at 59.04 GPixel/s versus 27.44 GPixel/s, and texture rate favors it at 162.4 GTexel/s versus 109.8 GTexel/s.