AMD Radeon RX 590 vs AMD Radeon RX 6700 XT Comparison

Head-to-Head Benchmarks

The recorded database contains two direct head-to-head comparisons between the AMD Radeon RX 6700 XT and the AMD Radeon RX 590, and in both instances the RX 6700 XT secures a commanding victory. The first test, 3DMark Steel Nomad DX12, shows the RX 6700 XT scoring 2435 points against the RX 590's 1081 points. That represents a 125.3% advantage, meaning the newer card more than doubles the older card's performance in this particular DirectX 12 workload. Such a margin is not incremental; it is a generational leap that places the RX 6700 XT in a completely different performance tier.

The second head-to-head result, Geekbench Metal, is even more lopsided. The RX 6700 XT records 130896 points, while the RX 590 manages only 48407 points. The delta expands to 170.4%, which is a substantial increase over the DX12 gap. This is notable because Metal is Apple's graphics API, and the result indicates that the RX 6700 XT's architecture scales particularly well in compute-oriented or Metal-optimized workloads compared to the older GCN design. The RX 6700 XT wins both recorded head-to-head tests, giving it a 2-0 record over the RX 590.

Contextualizing these scores against their respective peer groups reinforces the separation. The RX 6700 XT's average benchmark score across all recorded tests is 27425, which places it in the 72nd percentile of all GPUs in the database. Its nearest rivals, according to average score, are the NVIDIA GeForce RTX 4070 Mobile at 27435, a negligible 0% delta, and the NVIDIA GeForce RTX 3090 at 27565, which is only 0.5% ahead. The RX 6700 XT also trails the AMD Radeon Pro Vega 20 by 1.5%, which scores 27839, and it leads the NVIDIA RTX PRO 4000 Blackwell by 1.1%. These delta values are all within a narrow band, meaning the RX 6700 XT sits in a tightly contested performance neighborhood despite its 72nd percentile ranking.

The RX 590, by contrast, has an average benchmark score of 24744, which places it in the 70th percentile of all GPUs. Its nearest rivals include the NVIDIA RTX A5000 Mobile at 24763, a 0.1% deficit, and the Intel Arc A350M at 24647, which trails by 0.4%. The RX 590 leads the AMD Radeon RX 6600 XT by 1.2% and the NVIDIA GeForce GTX 1630 by 1.9% in average score. The percentile gap between the two cards is modest, only 2 percentage points, but the raw average score difference is 2681 points in favor of the RX 6700 XT. That discrepancy between percentile and average score suggests the RX 590 benefits from a dense cluster of similarly performing GPUs, whereas the RX 6700 XT sits among faster competition.

Looking at the individual benchmark suite, the RX 6700 XT shows a broad spread of capabilities. Its PassMark G3D score is 19786, its PassMark G2D score is 951, and its PassMark GPU Compute score is 9336. In legacy DirectX tests, it scores 242 in DirectX 9, 173 in DirectX 11, and 124 in DirectX 10, with a DirectX 12 score of 77. The Geekbench OpenCL score is 44152, and the Geekbench Vulkan score is 93500. The RX 590's recorded benchmark data is much sparser, with only the two head-to-head tests listed, so a full multi-test comparison is not possible from the database. However, the two tests that do exist are sufficient to establish the performance hierarchy.

The Verdict

The data is unambiguous: the AMD Radeon RX 6700 XT is the faster GPU in every recorded comparison. Its 125.3% lead in 3DMark Steel Nomad DX12 and 170.4% lead in Geekbench Metal leave no room for interpretation. For users prioritizing raw performance in modern DirectX 12 workloads or Metal-based applications, the RX 6700 XT is the clear choice. The RX 590, while still holding a respectable 70th percentile ranking, operates in a different performance class entirely.

The RX 6700 XT also brings a more modern feature set that matters for certain workloads. It supports DirectX 12 Ultimate with feature level 12_2, whereas the RX 590 is limited to DirectX 12 with feature level 12_0. The RX 6700 XT includes 40 ray tracing cores, a capability the RX 590 lacks entirely. Its Vulkan support is version 1.4 compared to the RX 590's Vulkan 1.3. These are not just specification differences; they translate into real functional gaps for applications that leverage ray tracing or newer API features.

For users who do not need those newer features and are constrained by other factors, the RX 590 still has a role. Its average benchmark score of 24744 places it in the 70th percentile, and its nearest rivals include the NVIDIA GeForce GTX 1630, which it leads by 1.9%. It is a dual-slot card with a 175 W TDP and requires only a 450 W suggested PSU, making it a lighter load on system power delivery. The RX 6700 XT, by contrast, has a 230 W TDP and a 550 W suggested PSU. The RX 590 also uses a single 8-pin power connector, while the RX 6700 XT needs one 6-pin and one 8-pin connector.

The decision, strictly from the data, comes down to performance needs versus system constraints. The RX 6700 XT delivers more than double the performance in the recorded head-to-head tests and adds ray tracing support, newer API versions, and a higher memory bandwidth. The RX 590 offers lower power requirements and a smaller physical footprint, with a length of 241 mm compared to the RX 6700 XT's 267 mm. Neither card is currently in production, as both are marked end-of-life, but the RX 6700 XT is the superior choice for anyone seeking maximum performance from this pairing.

Architecture Differences

The two cards represent two distinct eras of AMD GPU design. The RX 6700 XT is built on the RDNA 2.0 architecture, using the Navi 22 chip, and belongs to the Navi II generation from the RX 6000 series. The RX 590 uses the older GCN 4.0 architecture, built on the Polaris 30 chip, from the Polaris generation in the RX 500 series. This architectural divide explains the substantial performance differences recorded in the benchmarks.

The manufacturing process underscores the generational gap. The RX 6700 XT is fabricated on a 7 nm process at TSMC, while the RX 590 uses a 12 nm process at GlobalFoundries. Transistor counts reflect this difference: the RX 6700 XT packs 17,200 million transistors on a 335 mm² die, yielding a transistor density of 51.3 million per square millimeter. The RX 590 contains 5,700 million transistors on a 232 mm² die, with a density of 24.6 million per square millimeter. The RX 6700 XT crams roughly three times as many transistors into only 44% more die area, a direct consequence of the denser 7 nm process.

Core configuration also differs significantly. The RX 6700 XT has 2560 shading units, 160 texture mapping units, and 64 render output units, alongside 40 dedicated ray tracing cores. The RX 590 has 2304 shading units, 144 TMUs, and only 32 ROPs, with no ray tracing hardware. The shading unit count is only 11% higher on the RX 6700 XT, but the ROP count is exactly double, which contributes to its much higher pixel throughput. The recorded pixel rate for the RX 6700 XT is 165.2 GPixel/s against the RX 590's 49.44 GPixel/s, a 3.3x advantage. Texture rate tells a similar story: 413.0 GTexel/s versus 222.5 GTexel/s, a 1.9x lead.

Compute performance shows the same pattern. The RX 6700 XT delivers 13.21 TFLOPS of FP32 performance, while the RX 590 delivers 7.119 TFLOPS. The RX 6700 XT is 85% faster in raw single-precision floating point. FP16 performance diverges even more sharply because of how each architecture handles half-precision. The RX 6700 XT achieves 26.43 TFLOPS with a 2:1 FP16 rate, effectively doubling its FP32 throughput. The RX 590 manages only 7.119 TFLOPS in FP16 at a 1:1 rate, meaning it gains nothing from half-precision. This makes the RX 6700 XT nearly 3.7x faster in FP16 workloads, a significant advantage for compute tasks that leverage reduced precision.

Clock speeds also favor the RX 6700 XT. Its base clock is 2321 MHz, its boost clock is 2581 MHz, and it has a game clock of 2424 MHz. The RX 590's base clock is 1469 MHz with a boost clock of 1545 MHz and no recorded game clock. The RX 6700 XT's boost clock is 67% higher than the RX 590's, which compounds with the architectural efficiency gains of RDNA 2.0. The memory clock is listed as 2000 MHz for both cards, but the effective data rates differ: 16 Gbps for the RX 6700 XT versus 8 Gbps for the RX 590. That doubling of effective memory speed, combined with the different memory types, explains the bandwidth gap.

Specification Differences

Memory configuration is one of the most visible differences. The RX 6700 XT has 12 GB of GDDR6 memory on a 192-bit bus, delivering 384.0 GB/s of bandwidth. The RX 590 has 8 GB of GDDR5 memory on a 256-bit bus, delivering 256.0 GB/s. Despite the RX 590's wider bus, the RX 6700 XT's faster GDDR6 memory gives it 50% more bandwidth. The RX 6700 XT also has 4 GB more memory capacity, which matters for texture-heavy workloads or larger frame buffers.

Power and cooling requirements differ substantially. The RX 6700 XT has a TDP of 230 W and requires a 550 W suggested PSU, along with one 6-pin and one 8-pin power connector. The RX 590 has a TDP of 175 W, a 450 W suggested PSU, and a single 8-pin connector. Both cards are dual-slot designs. Physical dimensions differ, with the RX 6700 XT measuring 267 mm in length, 110 mm in height, and 40 mm in width, while the RX 590 measures 241 mm in length with no recorded height or width values.

The bus interface also differs. The RX 6700 XT uses PCIe 4.0 x16, while the RX 590 uses PCIe 3.0 x16. This has implications for data transfer rates between the GPU and the rest of the system, though the practical impact depends on the workload and platform. Display outputs are similar in count but differ in version: the RX 6700 XT offers 1x HDMI 2.1 and 3x DisplayPort 1.4a, while the RX 590 offers 1x HDMI 2.0b and 3x DisplayPort 1.4a. The newer HDMI standard on the RX 6700 XT supports higher resolutions and refresh rates over HDMI connections.

API support reflects the architectural differences. The RX 6700 XT supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The RX 590 supports DirectX 12 (12_0), OpenGL 4.6, and Vulkan 1.3. The DirectX feature level difference is particularly important for ray tracing and mesh shaders, which require 12_2 support. The RX 6700 XT's ray tracing cores are the physical hardware that enables this capability, and the RX 590 has none.

Release timing and pricing are also recorded. The RX 6700 XT was released on 2021-03-02 with a launch MSRP of 479 USD. The RX 590 was released on 2018-11-14 with a launch MSRP of 279 USD. Both are end-of-life products. The RX 6700 XT's predecessor is listed as Navi, and its successor is Navi III. The RX 590's predecessor is Arctic Islands, and its successor is Vega.

FAQ

Q: Which GPU is faster in the recorded head-to-head benchmarks?

A: The AMD Radeon RX 6700 XT wins both recorded tests. It scores 2435 versus 1081 in 3DMark Steel Nomad DX12, a 125.3% lead, and 130896 versus 48407 in Geekbench Metal, a 170.4% lead.

Q: Does the RX 6700 XT support ray tracing?

A: Yes. The RX 6700 XT includes 40 ray tracing cores and supports DirectX 12 Ultimate with feature level 12_2. The RX 590 has no ray tracing cores and is limited to DirectX 12 with feature level 12_0.

Q: What are the memory specifications of each card?

A: The RX 6700 XT has 12 GB of GDDR6 memory on a 192-bit bus with 384.0 GB/s bandwidth. The RX 590 has 8 GB of GDDR5 memory on a 256-bit bus with 256.0 GB/s bandwidth.

Q: How do their power requirements compare?

A: The RX 6700 XT has a 230 W TDP and requires a 550 W suggested PSU with one 6-pin and one 8-pin connector. The RX 590 has a 175 W TDP and requires a 450 W suggested PSU with a single 8-pin connector.

Q: Which card has better compute performance in FP16?

A: The RX 6700 XT delivers 26.43 TFLOPS FP16 with a 2:1 ratio, while the RX 590 delivers 7.119 TFLOPS FP16 at a 1:1 ratio. The RX 6700 XT is approximately 3.7x faster in half-precision workloads.

Q: What are the average benchmark scores and percentile rankings for each card?

A: The RX 6700 XT has an average benchmark score of 27425 and ranks in the 72nd percentile of all GPUs. The RX 590 has an average benchmark score of 24744 and ranks in the 70th percentile.