AMD Radeon RX 6900 XT vs NVIDIA GeForce RTX 5080 Comparison

The data presents a decisive generational mismatch. The NVIDIA GeForce RTX 5080, built on the Blackwell 2.0 architecture, holds a commanding lead over the AMD Radeon RX 6900 XT across every single benchmark recorded in the head-to-head comparison. The RTX 5080 wins all 10 tests, with an average benchmark score of 56,083 compared to the RX 6900 XT's 50,951. This is not a close contest; it is a demonstration of how far GPU technology has advanced between the 2020 RDNA 2.0 flagship and the 2025 Blackwell release. The RTX 5080 also carries a higher percentile ranking among all GPUs at 87, versus the RX 6900 XT's 86, though the latter's position is still respectable given its age. The RX 6900 XT, now marked as end-of-life, remains a capable card, but the benchmark results show it cannot match the newer NVIDIA product in any measurable way.

The Verdict

The RTX 5080 is the clear winner for any user prioritizing raw performance. Its average benchmark score of 56,083 is 10.1% higher than the RX 6900 XT's 50,951, and it wins every single direct comparison. The largest victory comes in the 3DMark Steel Nomad DX12 test, where the RTX 5080 scores 8,637 against the RX 6900 XT's 4,079—a 111.7% advantage. This indicates that for modern, demanding DX12 workloads, the RTX 5080 is more than twice as fast. The verdict is straightforward: the RTX 5080 is the superior product for gaming and compute, and the RX 6900 XT should only be considered if found at a significant discount in the used market, though pricing data is not part of this analysis.

The RX 6900 XT's only consolation is its lower power draw of 300 W versus the RTX 5080's 360 W, and its triple-slot cooler suggests it was designed for high thermal loads. However, in terms of performance, it is outclassed. The RTX 5080 also offers a more modern feature set, including PCIe 5.0 x16 interface and DisplayPort 2.1b outputs, while the RX 6900 XT is limited to PCIe 4.0 and DisplayPort 1.4a. For a user building a new system today, the RTX 5080 is the logical choice. For someone with an existing RX 6900 XT, the data shows an upgrade to the RTX 5080 would yield substantial gains across all workloads, but whether that justifies the cost is a separate question outside the scope of these benchmarks.

Where Each One Wins

The RTX 5080 wins in every single benchmark category, leaving no area where the RX 6900 XT can claim a victory. The most dramatic wins are in compute-heavy and modern API tests. In 3DMark Steel Nomad DX12, the RTX 5080's 8,637 score is more than double the RX 6900 XT's 4,079. This suggests the RTX 5080 is significantly better optimized for next-generation game engines that leverage DX12's advanced features. Similarly, in Geekbench Vulkan, the RTX 5080 scores 255,450 versus 148,525, a 72% advantage, indicating strong performance in cross-platform graphics APIs.

In less modern workloads, the RTX 5080 still leads, but by smaller margins. For example, in PassMark DirectX 10, the RTX 5080 scores 208 against the RX 6900 XT's 167, a 24.6% lead. This shows that even in older APIs, the newer architecture holds an advantage. The RTX 5080 also wins in compute tasks, with PassMark GPU Compute scores of 21,789 versus 14,547, a 49.8% lead. This makes the RTX 5080 the better choice for GPU-accelerated compute workloads like rendering or scientific simulation. The RX 6900 XT, with its 16 GB of GDDR6 memory, offers the same memory capacity as the RTX 5080, but the RTX 5080's GDDR7 memory provides significantly higher bandwidth. In summary, the RTX 5080 is the winner for every use case, from legacy DirectX 9 titles to the latest DX12 and Vulkan games, as well as compute-heavy applications.

Architecture Differences

The architectural gap between these two GPUs is vast. The RTX 5080 is built on a 5 nm process at TSMC, while the RX 6900 XT uses a 7 nm process. This allows the RTX 5080 to pack 45,600 million transistors into a 378 mm² die, resulting in a transistor density of 120.6 million per mm². In contrast, the RX 6900 XT has 26,800 million transistors on a much larger 520 mm² die, with a density of just 51.5 million per mm². This explains how the RTX 5080 achieves higher performance despite being a physically smaller chip.

The RTX 5080's Blackwell 2.0 architecture is a major evolution over the RX 6900 XT's RDNA 2.0. The RTX 5080 features 10,752 shading units, 336 texture mapping units, and 112 ROPs, compared to the RX 6900 XT's 5,120 shading units, 320 TMUs, and 128 ROPs. The RTX 5080 also has more ray tracing cores (84 vs 80) and adds 336 tensor cores, which the RX 6900 XT lacks entirely. These tensor cores are crucial for AI-accelerated features like DLSS, which the RX 6900 XT cannot leverage. The RTX 5080's FP32 compute performance is 56.28 TFLOPS, more than double the RX 6900 XT's 23.04 TFLOPS. This raw compute advantage is reflected in the benchmark results.

Memory architecture also differs significantly. The RTX 5080 uses 16 GB of GDDR7 memory on a 256-bit bus, delivering 960.0 GB/s of bandwidth. The RX 6900 XT also has 16 GB, but it is GDDR6 on a 256-bit bus, providing only 512.0 GB/s. This 87.5% bandwidth advantage for the RTX 5080 is critical for high-resolution textures and compute workloads. The RTX 5080 also uses a PCIe 5.0 x16 interface, while the RX 6900 XT is limited to PCIe 4.0 x16, doubling the potential data transfer rate to the CPU. Finally, the RTX 5080 supports DisplayPort 2.1b outputs, whereas the RX 6900 XT has DisplayPort 1.4a, which means the newer card can drive higher refresh rates at higher resolutions on compatible monitors.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The NVIDIA GeForce RTX 5080 has an average benchmark score of 56,083, while the AMD Radeon RX 6900 XT has an average of 50,951. This represents a 10.1% advantage for the RTX 5080.

Q: How does the RTX 5080 perform in DirectX 12 compared to the RX 6900 XT?

A: In the 3DMark Steel Nomad DX12 test, the RTX 5080 scores 8,637, which is 111.7% higher than the RX 6900 XT's 4,079. This shows the RTX 5080 is more than twice as fast in this modern DX12 workload.

Q: Does the RX 6900 XT have any tensor cores?

A: No, the RX 6900 XT does not have any tensor cores. The RTX 5080 has 336 tensor cores, which enable AI-accelerated features that the AMD card cannot perform.

Q: What is the memory bandwidth difference between the two cards?

A: The RTX 5080 has a memory bandwidth of 960.0 GB/s, while the RX 6900 XT has 512.0 GB/s. This is a 87.5% advantage for the RTX 5080.

Q: Which GPU is more power-efficient based on the data?

A: The RX 6900 XT has a lower TDP of 300 W, compared to the RTX 5080's 360 W. However, the RTX 5080 delivers significantly higher performance, so efficiency per watt is not directly comparable from this data alone.

Q: Is the RX 6900 XT still in production?

A: No, the RX 6900 XT is listed as end-of-life, while the RTX 5080 is still an active product. This means the RTX 5080 is more readily available and supported.

Head-to-Head Benchmarks

The head-to-head benchmark results are unequivocal: the RTX 5080 wins all 10 tests, with victories ranging from 16.1% to 111.7%. The most lopsided result is in 3DMark Steel Nomad DX12, where the RTX 5080 scores 8,637 versus the RX 6900 XT's 4,079. This 111.7% delta is a massive leap and indicates that the RTX 5080 is exceptionally well-suited for next-generation games using DX12 features like mesh shaders and ray tracing.

In Geekbench Vulkan, the RTX 5080 scores 255,450, a 72% advantage over the RX 6900 XT's 148,525. This suggests the RTX 5080 is also superior in cross-platform graphics workloads, which is important for games that run on multiple operating systems. The Geekbench OpenCL test shows a smaller but still significant lead of 25.7%, with scores of 235,901 for the RTX 5080 and 187,673 for the RX 6900 XT. This indicates better general-purpose compute performance, which is relevant for non-gaming tasks.

In the PassMark suite, the RTX 5080 leads in every category. The largest PassMark win is in GPU Compute, where the RTX 5080 scores 21,789 versus 14,547, a 49.8% advantage. This is a critical result for users who rely on GPU acceleration for tasks like video encoding or machine learning. In PassMark G3D, the RTX 5080 scores 36,565 against 26,732, a 36.8% lead. Even in older DirectX 9 and 10 tests, the RTX 5080 wins by 45.1% and 24.6% respectively, showing that the newer architecture is more efficient across all API generations. The smallest win is in PassMark DirectX 11, where the RTX 5080 leads by 16.1%, but it still wins. These results paint a clear picture: the RTX 5080 is comprehensively faster than the RX 6900 XT in every benchmark scenario.

Specification Differences

The key specification differences between the two GPUs highlight the generational leap. The RTX 5080 is built on a 5 nm process, while the RX 6900 XT uses 7 nm. The RTX 5080 has 45,600 million transistors on a 378 mm² die, whereas the RX 6900 XT has 26,800 million on a 520 mm² die. The RTX 5080's base clock is 2295 MHz and boost clock is 2617 MHz, while the RX 6900 XT has a base clock of 1825 MHz and boost of 2250 MHz. The RTX 5080's memory runs at 1875 MHz (30 Gbps effective), while the RX 6900 XT's memory runs at 2000 MHz (16 Gbps effective).

The RTX 5080 has 10,752 shading units, 336 TMUs, and 112 ROPs, while the RX 6900 XT has 5,120 shading units, 320 TMUs, and 128 ROPs. The RTX 5080 has 84 ray tracing cores and 336 tensor cores, while the RX 6900 XT has 80 ray tracing cores and no tensor cores. The RTX 5080's FP32 performance is 56.28 TFLOPS, while the RX 6900 XT's is 23.04 TFLOPS. Its FP16 performance is 56.28 TFLOPS (1:1), while the RX 6900 XT's is 46.08 TFLOPS (2:1). The RTX 5080 has a TDP of 360 W and uses a single 16-pin power connector, while the RX 6900 XT has a TDP of 300 W and uses two 8-pin connectors. The RTX 5080 is dual-slot, while the RX 6900 XT is triple-slot. The RTX 5080 uses PCIe 5.0 x16, while the RX 6900 XT uses PCIe 4.0 x16. Finally, the RTX 5080 has 1x HDMI 2.1b and 3x DisplayPort 2.1b outputs, while the RX 6900 XT has 1x HDMI 2.1, 2x DisplayPort 1.4a, and 1x USB Type-C.