AMD Radeon RX 6650M XT vs NVIDIA GeForce RTX 5080 Comparison

Where Each One Wins

The benchmark data paints a decisively one-sided picture in this comparison. The NVIDIA GeForce RTX 5080 wins the only head-to-head test recorded in the database, the Geekbench OpenCL compute benchmark, with a score of 235,901 against the AMD Radeon RX 6650M XT's 76,904. That translates to a 67.4% deficit for the AMD part, meaning the RTX 5080 delivers roughly three times the raw compute throughput in that workload.

For the AMD Radeon RX 6650M XT, the wins are not found in performance supremacy but in its positioning as a mobile-first product. The database shows it carries a 120 W TDP, uses no external power connectors, and is designed as an integrated graphics processor (IGP) form factor intended for portable devices. Its display outputs are listed as "Portable Device Dependent," underscoring that this is a part built for laptops rather than desktop towers. The RTX 5080, by contrast, is a dual-slot desktop card requiring a 360 W TDP, a single 16-pin power connector, and a suggested 750 W power supply. So the AMD part "wins" in the sense of being a self-contained mobile solution that fits into thin-and-light chassis without external power cabling, while the NVIDIA part is a high-power desktop behemoth.

The RTX 5080 also sweeps the broader benchmark suite. Beyond the OpenCL head-to-head, it records a Vulkan score of 255,450, a Passmark G3D score of 36,565, and a Passmark GPU Compute score of 21,789. The AMD RX 6650M XT has no recorded scores for those tests in the database, so its only data point remains the 76,904 OpenCL result. In terms of wins across the entire database, the tally is 0 for AMD and 1 for NVIDIA in the direct head-to-head, but the RTX 5080's additional benchmark entries give it a far richer performance profile.

The percentile rankings are telling as well. The AMD part sits at the 91st percentile among all GPUs, which sounds strong, but that is based on a single OpenCL score. The RTX 5080 sits at the 87th percentile, yet its average benchmark score across ten tests is 56,083, which is lower than the AMD part's 76,904 average. This discrepancy arises because the RTX 5080's average includes low-scoring Passmark DX9, DX10, DX11, and DX12 results (389, 208, 324, and 151 respectively), which drag down its mean. The AMD part's average is inflated by having only one high-scoring entry. In practical terms, the RTX 5080 is the clear performance leader in every compute and graphics workload where both have data.

FAQ

Q: Which GPU wins the only direct head-to-head benchmark in the database?

A: The NVIDIA GeForce RTX 5080 wins the Geekbench OpenCL test with a score of 235,901 versus 76,904 for the AMD Radeon RX 6650M XT, a delta of -67.4% for the AMD part.

Q: How do the memory specifications differ between the two cards?

A: The AMD Radeon RX 6650M XT has 8 GB of GDDR6 memory on a 128-bit bus with 256.0 GB/s bandwidth. The NVIDIA GeForce RTX 5080 has 16 GB of GDDR7 memory on a 256-bit bus with 960.0 GB/s bandwidth.

Q: What is the power requirement difference?

A: The AMD part has a 120 W TDP and uses no power connectors, making it suitable for portable devices. The NVIDIA part has a 360 W TDP, requires a single 16-pin power connector, and the database lists a suggested power supply of 750 W.

Q: Are both GPUs based on the same manufacturing process?

A: No. The AMD Radeon RX 6650M XT uses a 7 nm process from TSMC, while the NVIDIA GeForce RTX 5080 uses a 5 nm process, also from TSMC.

Q: What is the release timeline for these products?

A: The AMD Radeon RX 6650M XT was released on January 3, 2022, and is marked as end-of-life. The NVIDIA GeForce RTX 5080 was released on January 29, 2025, and is listed as active production.

Q: Does the RTX 5080 have tensor cores?

A: Yes, the NVIDIA GeForce RTX 5080 includes 336 tensor cores. The AMD Radeon RX 6650M XT has no tensor cores listed; instead it has 32 ray tracing cores.

Head-to-Head Benchmarks

The database records exactly one head-to-head benchmark between these two GPUs: Geekbench OpenCL. The NVIDIA GeForce RTX 5080 scores 235,901, while the AMD Radeon RX 6650M XT scores 76,904. The delta percentage is -67.4%, meaning the AMD part achieves less than a third of the NVIDIA card's score. This is a massive gap in raw compute throughput, reflecting the generational and architectural differences between the two products.

Looking at the RTX 5080's other recorded benchmarks, the pattern of dominance continues. Its Geekbench Vulkan score is 255,450, which is even higher than its OpenCL result. In Passmark tests, it scores 36,565 in G3D (the main 3D graphics benchmark) and 21,789 in GPU Compute. The DirectX legacy tests are lower, with DX9 at 389, DX10 at 208, DX11 at 324, and DX12 at 151, but those are older API workloads where the card's modern architecture may not be fully utilized. The 2D test score is 1,415.

For the AMD Radeon RX 6650M XT, the only recorded benchmark is the OpenCL score of 76,904. There are no Vulkan, Passmark, or 3DMark entries in the database for this part. Its nearest rivals in the database give context: the NVIDIA GeForce RTX 5090 D scores 77,712 (1% higher), the AMD Radeon RX 6850M XT scores 78,940 (2.6% higher), and the NVIDIA Tesla P100 variants score 79,396 and 79,605 (3.1% and 3.4% higher respectively). So the RX 6650M XT sits in a cluster of GPUs around the 77,000 to 80,000 OpenCL range, all closely matched within a few percentage points.

The RTX 5080's nearest rivals tell a different story. Its average benchmark score of 56,083 is close to the AMD Radeon 8060S (55,757, 0.6% lower) and the AMD Radeon RX 6750 GRE 12 GB (55,698, 0.7% lower). The AMD Radeon Pro W5700X is 2.3% lower, and the AMD Radeon RX 9070 GRE is 2.2% higher. Note that these rivals are compared on average benchmark score, not the OpenCL head-to-head, so they reflect a different aggregate metric.

The 3DMark Steel Nomad DX12 test for the RTX 5080 records a score of 8,637, which is a modern DirectX 12 workload. The AMD part has no corresponding entry. In every measurable dimension, the RTX 5080 outperforms the RX 6650M XT by a wide margin, often by multiples.

Specification Differences

The two GPUs differ in nearly every specification category. Memory is a clear differentiator: the AMD part has 8 GB of GDDR6 on a 128-bit bus with 256.0 GB/s bandwidth, while the NVIDIA part has 16 GB of GDDR7 on a 256-bit bus with 960.0 GB/s bandwidth. The memory clock also differs, with the AMD part running at 2000 MHz (16 Gbps effective) and the NVIDIA part at 1875 MHz (30 Gbps effective), showing that the GDDR7 standard delivers far higher effective data rates despite a lower base clock.

Compute resources are massively different. The AMD Radeon RX 6650M XT has 2,048 shading units, 128 texture mapping units, and 64 raster operation pipelines. The NVIDIA GeForce RTX 5080 has 10,752 shading units, 336 TMUs, and 112 ROPs. The RTX 5080 also has 84 ray tracing cores and 336 tensor cores, while the AMD part has 32 ray tracing cores and no tensor cores listed. Pixel rate is 154.6 GPixel/s for AMD versus 293.1 GPixel/s for NVIDIA. Texture rate is 309.2 GTexel/s versus 879.3 GTexel/s. FP32 performance is 9.896 TFLOPS versus 56.28 TFLOPS, a 5.7x difference. FP16 performance is 19.79 TFLOPS (2:1 ratio) for AMD versus 56.28 TFLOPS (1:1 ratio) for NVIDIA, meaning the NVIDIA card does not rely on a packed math trick to reach its FP16 number.

Clock speeds also differ. The AMD part has a base clock of 2068 MHz, a boost clock of 2416 MHz, and a game clock of 2162 MHz. The NVIDIA part has a base clock of 2295 MHz and a boost clock of 2617 MHz, with no game clock listed. Power and physical specs are starkly different: 120 W TDP versus 360 W TDP, IGP slot width versus dual-slot, no power connectors versus a single 16-pin connector, and no suggested PSU versus a 750 W suggestion. The bus interface is PCIe 4.0 x8 for AMD and PCIe 5.0 x16 for NVIDIA. Display outputs are "Portable Device Dependent" for the AMD part versus 1x HDMI 2.1b and 3x DisplayPort 2.1b for the NVIDIA part. The NVIDIA card has physical dimensions of 304 mm length, 137 mm height, and 40 mm width; the AMD part has no listed dimensions.

Architecture Differences

The architectural gap is generational. The AMD Radeon RX 6650M XT is built on RDNA 2.0, using the Navi 23 chip, and belongs to the Radeon RX 6000 series from the Navi Mobile (RX 6000M) generation. The NVIDIA GeForce RTX 5080 is built on Blackwell 2.0, using the GB203 chip, and belongs to the GeForce 50-series. This is a comparison between a 2022 mobile part and a 2025 desktop part.

The manufacturing process differs: the AMD chip is on a 7 nm TSMC node, while the NVIDIA chip is on a 5 nm TSMC node. Transistor counts reflect the scale difference: the AMD Navi 23 has 11,060 million transistors on a 237 mm² die, giving a transistor density of 46.7 million per mm². The NVIDIA GB203 has 45,600 million transistors on a 378 mm² die, giving a density of 120.6 million per mm². That is a 4.1x transistor count increase and a 2.6x density increase, showing both a larger chip and a more advanced process.

The ray tracing and compute feature sets are also different. The AMD part has 32 dedicated ray tracing cores, which is the RDNA 2 approach to hardware-accelerated ray tracing. The NVIDIA part has 84 ray tracing cores plus 336 tensor cores, the latter enabling AI-accelerated features like DLSS that the AMD part cannot match. The AMD part has no tensor core equivalent. The FP16 execution ratio is another architectural tell: AMD uses a 2:1 packed rate (19.79 TFLOPS from 9.896 TFLOPS FP32), while NVIDIA runs FP16 at a full 1:1 rate (56.28 TFLOPS), meaning every FP32 unit also executes FP16 without a packed instruction path.

The API support is identical on paper: both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The production status differs, with the AMD part marked end-of-life and the NVIDIA part active. The AMD part's predecessor is Polaris Mobile, while the NVIDIA part's predecessor is GeForce 40 and its successor is GeForce 60, according to the database. The release dates are January 2022 for AMD and January 2025 for NVIDIA, a three-year gap that explains the architectural leap. The launch MSRP for the RTX 5080 is 999 USD, while the AMD part has no listed launch MSRP.