AMD Radeon RX 6900 XT vs NVIDIA GeForce RTX 5090 Mobile Comparison

AMD Radeon RX 6900 XT vs NVIDIA GeForce RTX 5090 Mobile is a matchup between a 2020 desktop flagship and a 2025 mobile part, and the data reflects a generational and form-factor divide. The RX 6900 XT holds a higher aggregate benchmark score and percentile ranking, but the RTX 5090 Mobile wins the majority of individual tests, including the modern 3DMark Steel Nomad workload by a wide margin. This analysis compares the two based solely on the supplied benchmark scores, architectural specifications, and nearest-rival context.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The AMD Radeon RX 6900 XT has an average benchmark score of 50,951, while the NVIDIA GeForce RTX 5090 Mobile scores 45,152. The AMD part also sits in the 86th percentile of all GPUs, two points above the RTX 5090 Mobile’s 84th percentile.

Q: How much faster is the RTX 5090 Mobile in 3DMark Steel Nomad?

A: The RTX 5090 Mobile scores 5,871 in 3DMark Steel Nomad, which is 30.5% higher than the RX 6900 XT’s 4,079. This is the largest single-test delta in the head-to-head results.

Q: Does the RTX 5090 Mobile win every benchmark?

A: No. The RX 6900 XT wins two of the ten head-to-head tests: Passmark DirectX 11 with a 3.7% lead (279 vs 269) and Passmark GPU Compute with an 8.6% lead (14,547 vs 13,401). The RTX 5090 Mobile wins the remaining eight.

Q: What is the memory configuration difference?

A: The RX 6900 XT has 16 GB of GDDR6 on a 256-bit bus with 512.0 GB/s bandwidth. The RTX 5090 Mobile has 24 GB of GDDR7 on the same 256-bit bus width but delivers 896.0 GB/s bandwidth.

Q: Which GPU has a higher transistor density?

A: The RTX 5090 Mobile uses a 5 nm process with a transistor density of 120.6M / mm². The RX 6900 XT uses a 7 nm process with a density of 51.5M / mm². The mobile chip packs 45,600 million transistors on a 378 mm² die, versus 26,800 million on a 520 mm² die for the desktop part.

Q: What is the TDP difference?

A: The RX 6900 XT has a 300 W TDP and requires a 700 W suggested PSU with 2x 8-pin connectors. The RTX 5090 Mobile has a 95 W TDP, uses no external power connectors, and is an integrated graphics package (IGP).

The Verdict

The data paints a clear picture: the RTX 5090 Mobile is the faster GPU in most modern workloads, despite being a mobile part with a 95 W TDP. It wins eight of ten head-to-head benchmarks, with decisive victories in 3DMark Steel Nomad (30.5% ahead), Geekbench Vulkan (25.1% ahead), and Passmark DirectX 12 (18.1% ahead). For anyone prioritizing current-generation gaming performance, ray tracing, or compute via Vulkan, the RTX 5090 Mobile is the superior choice based on these scores.

However, the RX 6900 XT still holds advantages in specific legacy and compute scenarios. It wins Passmark DirectX 11 (3.7% ahead) and Passmark GPU Compute (8.6% ahead). Its average benchmark score is 12.8% higher than the RTX 5090 Mobile’s, and its 86th percentile ranking beats the mobile part’s 84th percentile. That makes the RX 6900 XT the better pick for users targeting older DirectX 11 titles or OpenCL-heavy workloads, where its higher raw FP32 throughput (23.04 TFLOPS vs 31.80 TFLOPS) apparently does not translate into a win in every compute test.

For a desktop builder with power budget to spare, the RX 6900 XT remains a capable end-of-life part. For a laptop buyer or anyone constrained to 95 W, the RTX 5090 Mobile is the stronger performer in current API benchmarks. The choice hinges on whether the user values the aggregate score and legacy wins of the desktop card, or the modern API dominance and mobile form factor of the RTX 5090 Mobile.

Head-to-Head Benchmarks

The RTX 5090 Mobile dominates the modern DirectX 12 and Vulkan workloads. In 3DMark Steel Nomad, it scores 5,871 against the RX 6900 XT’s 4,079, a 30.5% lead that represents the largest gap in the dataset. Geekbench Vulkan shows a similar trend: 198,405 vs 148,525, a 25.1% advantage. Geekbench OpenCL also favors the mobile part, with 201,834 vs 187,673, a 7% margin. These results indicate that the RTX 5090 Mobile’s Blackwell architecture is substantially better optimized for the latest graphics APIs.

The legacy DirectX tests are more split. The RTX 5090 Mobile wins Passmark DirectX 10 (183 vs 167, 8.7% ahead), DirectX 12 (138 vs 113, 18.1% ahead), and DirectX 9 (324 vs 268, 17.3% ahead). The RX 6900 XT claims only Passmark DirectX 11, scoring 279 vs 269, a 3.7% edge. This suggests the older AMD architecture retains a slight advantage in that specific driver path, but it is an isolated win.

In overall graphics performance, the RTX 5090 Mobile takes Passmark G3D with 30,034 vs 26,732, an 11% lead. The Passmark G2D result is nearly identical, with 1,057 vs 1,055 favoring the mobile part by a negligible 0.2%. The only other RX 6900 XT win is Passmark GPU Compute, where it scores 14,547 vs 13,401, an 8.6% advantage. That compute win is notable because the RTX 5090 Mobile has more shading units (10,496 vs 5,120) and higher FP32 peak (31.80 TFLOPS vs 23.04 TFLOPS), yet the AMD part still outperforms it in this specific test.

Overall, the head-to-head results show a 8-2 win count in favor of the RTX 5090 Mobile. The RX 6900 XT’s wins are concentrated in legacy DirectX 11 and one compute workload, while the RTX 5090 Mobile sweeps the modern API tests and most rasterization benchmarks.

Specification Differences

The two GPUs differ across nearly every physical and functional specification. The RX 6900 XT uses a 7 nm TSMC process with a 520 mm² die, while the RTX 5090 Mobile uses a 5 nm TSMC process with a 378 mm² die. Transistor counts are 26,800 million for AMD and 45,600 million for NVIDIA. The RTX 5090 Mobile achieves a much higher transistor density at 120.6M / mm² versus 51.5M / mm².

Clock speeds are starkly different. The RX 6900 XT runs at a base of 1825 MHz, a boost of 2250 MHz, and a game clock of 2015 MHz. The RTX 5090 Mobile runs at a base of 990 MHz and a boost of 1515 MHz, with no game clock listed. Memory clocks also differ: 2000 MHz (16 Gbps effective) for AMD versus 1750 MHz (28 Gbps effective) for NVIDIA.

The RTX 5090 Mobile has more shading units (10,496 vs 5,120), more TMUs (328 vs 320), and more RT cores (82 vs 80), but fewer ROPs (112 vs 128). The NVIDIA part also features 328 tensor cores, a capability the AMD card lacks entirely. Pixel rate favors AMD at 288.0 GPixel/s versus 169.7 GPixel/s, while texture rate favors AMD at 720.0 GTexel/s versus 496.9 GTexel/s. FP32 throughput is higher on NVIDIA at 31.80 TFLOPS versus 23.04 TFLOPS, but FP16 is lower on NVIDIA at 31.80 TFLOPS (1:1) versus 46.08 TFLOPS (2:1) on AMD.

Memory configurations are a significant differentiator. The RX 6900 XT has 16 GB of GDDR6 with 512.0 GB/s bandwidth, while the RTX 5090 Mobile has 24 GB of GDDR7 with 896.0 GB/s bandwidth, both on a 256-bit bus. The desktop card is triple-slot with 2x 8-pin power connectors and a 300 W TDP, while the mobile part is an IGP with no connectors and a 95 W TDP. The bus interface is PCIe 4.0 x16 for AMD and PCIe 5.0 x16 for NVIDIA. The RX 6900 XT has display outputs (1x HDMI 2.1, 2x DisplayPort 1.4a, 1x USB Type-C), while the RTX 5090 Mobile's outputs are listed as portable device dependent.

Architecture Differences

The RX 6900 XT is built on the RDNA 2.0 architecture (Navi 21 chip), part of the Navi II (RX 6000) generation. The RTX 5090 Mobile uses the Blackwell 2.0 architecture (GB203 chip), from the GeForce 50 Mobile generation. The process nodes are 7 nm for AMD and 5 nm for NVIDIA, both fabricated by TSMC.

Core organization differs fundamentally. The RX 6900 XT has 80 RT cores and no tensor cores, while the RTX 5090 Mobile has 82 RT cores and 328 tensor cores. This gives NVIDIA a dedicated hardware path for AI and DLSS workloads, which AMD lacks. Shading unit counts are nearly double on NVIDIA (10,496 vs 5,120), but AMD compensates with higher clock speeds and a wider texture pipeline (720.0 GTexel/s vs 496.9 GTexel/s).

The memory subsystem is also architecturally distinct. AMD uses GDDR6 with a 512.0 GB/s bandwidth, while NVIDIA uses GDDR7 with 896.0 GB/s bandwidth. Both share a 256-bit bus, but the mobile memory clock is lower (1750 MHz vs 2000 MHz) despite the higher effective data rate (28 Gbps vs 16 Gbps).

API support is identical: both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The production status differs, with the RX 6900 XT being end-of-life and the RTX 5090 Mobile active. Release dates are 2020-10-27 for AMD and 2025-03-26 for NVIDIA. The RX 6900 XT has a launch MSRP of 999 USD; the RTX 5090 Mobile has no listed launch MSRP. The RX 6900 XT’s predecessor is Navi and its successor is Navi III, while the RTX 5090 Mobile’s predecessor is GeForce 40 Mobile and it has no successor listed.