NVIDIA GeForce RTX 5080 vs NVIDIA GeForce RTX 5090 SE Comparison

Where Each One Wins

The recorded benchmark data splits cleanly between the two cards, but only one of them has any measured results in the database. The NVIDIA GeForce RTX 5080 has a full set of ten benchmark scores, while the NVIDIA GeForce RTX 5090 SE has no recorded benchmarks at all. This means every measured win belongs to the RTX 5080, simply because it is the only card in this comparison with test data.

The RTX 5080 shows its strength across a broad range of workloads. In 3DMark Steel Nomad DX12, it records a score of 8637. That test stresses modern gaming workloads with ray tracing and heavy geometry. The RTX 5080 also delivers strong compute results: Geekbench OpenCL at 235901 and Geekbench Vulkan at 255450. These are general-purpose compute scores that cover GPU acceleration in tasks such as rendering, simulation, and AI inference.

The Passmark suite adds more detail. The RTX 5080 scores 36565 in Passmark G3D, which is the aggregate gaming graphics score, and 21789 in Passmark GPU Compute, which isolates raw compute throughput. The older DirectX API tests show 389 in DirectX 9, 324 in DirectX 11, 208 in DirectX 10, and 151 in DirectX 12. The 2D graphics score is 1415. These results indicate the RTX 5080 handles legacy and modern APIs at different efficiency levels, with newer APIs not always producing the highest raw scores in this particular test suite.

Because the RTX 5090 SE has zero recorded benchmarks, the database cannot assign it any wins. Its average benchmark score is 0, and its percentile versus all GPUs sits at 50, which is the neutral midpoint used when no performance data exists. The RTX 5080, by contrast, has an average benchmark score of 56083 and sits at the 87th percentile versus all GPUs. That percentile places it above most of the database’s GPU population.

Architecture Differences

Both cards use the Blackwell 2.0 architecture and are built on TSMC’s 5 nm process node. Both belong to the GeForce 50-series and share the same DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4 API support. The similarities end at the chip level, where the two differ substantially.

The RTX 5080 uses the GB203 chip with 45,600 million transistors on a 378 mm² die, giving a transistor density of 120.6 million per square millimeter. The RTX 5090 SE uses the GB202 chip with 92,200 million transistors on a 750 mm² die, a density of 122.9 million per square millimeter. The GB202 has more than double the transistor count and nearly double the die area. That larger die provides the physical resources for more compute units.

Clock speeds differ significantly. The RTX 5080 has a base clock of 2295 MHz and a boost clock of 2617 MHz, while memory runs at 1875 MHz with 30 Gbps effective. The RTX 5090 SE has a lower base clock of 1740 MHz and a boost clock of 2377 MHz, with memory at 1750 MHz and 28 Gbps effective. The RTX 5080 clocks higher on the GPU core and memory, but the RTX 5090 SE compensates with far more execution units.

The memory subsystem is a major divergence. The RTX 5080 carries 16 GB of GDDR7 on a 256-bit bus with 960.0 GB/s bandwidth. The RTX 5090 SE carries 24 GB of GDDR7 on a 384-bit bus with 1.34 TB/s bandwidth. The wider bus and larger capacity give the RTX 5090 SE a clear advantage in memory-heavy workloads, even with slightly slower memory clocks.

Compute resources follow the chip size. The RTX 5080 has 10752 shading units, 336 texture mapping units, 112 render output units, 84 ray tracing cores, and 336 tensor cores. The RTX 5090 SE has 14080 shading units, 440 texture mapping units, 160 render output units, 110 ray tracing cores, and 440 tensor cores. Every count is higher on the RTX 5090 SE, with the largest relative gaps in render output units and tensor cores.

Pixel and texture rates reflect those resource differences. The RTX 5080 outputs 293.1 GPixel/s and 879.3 GTexel/s. The RTX 5090 SE outputs 380.3 GPixel/s and 1,045.9 GTexel/s. FP32 compute is 56.28 TFLOPS for the RTX 5080 and 66.94 TFLOPS for the RTX 5090 SE, with FP16 at 1:1 ratios for both. The RTX 5090 SE leads in every throughput metric despite lower clocks.

Head-to-Head Benchmarks

The database records no head-to-head benchmark entries between the RTX 5080 and the RTX 5090 SE. The head-to-head test list is empty, and the win counters for both cards are zero. This is an unusual situation for a comparison page, but the data does not support any direct matchup.

What the database does show is the RTX 5080’s full benchmark profile and its position relative to other GPUs. Its nearest rivals are all AMD cards. The AMD Radeon 8060S has an average score of 55757, which is 0.6% behind the RTX 5080’s 56083. The AMD Radeon RX 6750 GRE 12 GB scores 55698, trailing by 0.7%. The AMD Radeon Pro W5700X scores 54828, 2.3% behind. The AMD Radeon RX 9070 GRE scores 57367, which is 2.2% ahead of the RTX 5080.

These rival deltas give context for the RTX 5080’s standing. It sits in a tight cluster near 56,000 average points, with the closest rival only 0.6% away and the fastest rival 2.2% ahead. None of the nearest rivals are RTX 5090 SE cards. The RTX 5090 SE has no nearest rivals listed and no average score, so it cannot be placed on the same performance axis.

Without any recorded benchmarks for the RTX 5090 SE, the comparison must rely on architectural specifications to infer relative performance. The RTX 5090 SE has 31% more shading units than the RTX 5080 (14080 versus 10752), 31% more texture mapping units (440 versus 336), and 43% more render output units (160 versus 112). It also has 31% more ray tracing cores (110 versus 84) and 31% more tensor cores (440 versus 336). The FP32 throughput advantage of the RTX 5090 SE is 19% (66.94 TFLOPS versus 56.28 TFLOPS).

The memory bandwidth gap is larger. At 1.34 TB/s, the RTX 5090 SE provides 40% more bandwidth than the RTX 5080’s 960.0 GB/s. That bandwidth difference, combined with 50% more memory capacity (24 GB versus 16 GB), suggests the RTX 5090 SE would be substantially better in workloads that saturate memory, such as high-resolution texturing, large dataset processing, and certain compute tasks.

FAQ

Q: Which card has a higher boost clock?

A: The RTX 5080 boost clock is 2617 MHz, while the RTX 5090 SE boost clock is 2377 MHz. The RTX 5080 runs 240 MHz higher at boost.

Q: How much memory does each card have?

A: The RTX 5080 has 16 GB of GDDR7 on a 256-bit bus, while the RTX 5090 SE has 24 GB of GDDR7 on a 384-bit bus.

Q: What is the memory bandwidth for each card?

A: The RTX 5080 memory bandwidth is 960.0 GB/s. The RTX 5090 SE memory bandwidth is 1.34 TB/s, which is 40% higher.

Q: Which card has more ray tracing cores?

A: The RTX 5090 SE has 110 ray tracing cores. The RTX 5080 has 84, which is 26 fewer than the RTX 5090 SE.

Q: What are the power requirements for each card?

A: The RTX 5080 has a TDP of 360 W and a suggested PSU of 750 W. The RTX 5090 SE has a TDP of 500 W and a suggested PSU of 900 W.

Q: Does the RTX 5090 SE have any recorded benchmark scores?

A: No. The database lists an empty benchmark array for the RTX 5090 SE, with an average benchmark score of 0 and a percentile versus all GPUs of 50. The RTX 5080 has ten recorded benchmark scores and an average of 56083.

Specification Differences

| Specification | NVIDIA GeForce RTX 5080 | NVIDIA GeForce RTX 5090 SE |

|---|---|---|

| Chip | GB203 | GB202 |

| Transistors | 45,600 million | 92,200 million |

| Die Size | 378 mm² | 750 mm² |

| Transistor Density | 120.6M / mm² | 122.9M / mm² |

| Base Clock | 2295 MHz | 1740 MHz |

| Boost Clock | 2617 MHz | 2377 MHz |

| Memory Clock | 1875 MHz, 30 Gbps effective | 1750 MHz, 28 Gbps effective |

| Memory Size | 16 GB | 24 GB |

| Memory Bus Width | 256 bit | 384 bit |

| Memory Bandwidth | 960.0 GB/s | 1.34 TB/s |

| Shading Units | 10752 | 14080 |

| Texture Mapping Units | 336 | 440 |

| Render Output Units | 112 | 160 |

| Ray Tracing Cores | 84 | 110 |

| Tensor Cores | 336 | 440 |

| Pixel Rate | 293.1 GPixel/s | 380.3 GPixel/s |

| Texture Rate | 879.3 GTexel/s | 1,045.9 GTexel/s |

| FP32 Compute | 56.28 TFLOPS | 66.94 TFLOPS |

| FP16 Compute | 56.28 TFLOPS (1:1) | 66.94 TFLOPS (1:1) |

| TDP | 360 W | 500 W |

| Suggested PSU | 750 W | 900 W |

| Length | 304 mm, 12 inches | 267 mm, 10.5 inches |

| Height | 137 mm, 5.4 inches | 111 mm, 4.4 inches |

| Width | 40 mm, 1.6 inches | 40 mm, 1.6 inches |

| Launch MSRP | 999 USD | 1,499 USD |

| Release Date | 2025-01-29 | 2025-12-31 |

| Benchmark Count | 10 | 0 |

| Average Benchmark Score | 56083 | 0 |

| Percentile vs All GPUs | 87 | 50 |

The two cards share the same architecture, process node, foundry, PCIe interface, slot width, power connector, display outputs, and API support. They also have the same predecessor and successor generations. Every other measured specification differs, with the RTX 5090 SE leading in compute resources, memory capacity, bandwidth, physical die size, and power draw, while the RTX 5080 leads in clock speeds and physical compactness. The RTX 5090 SE is shorter and lower in height, making it a more compact card despite its larger die. The database cannot yet quantify the RTX 5090 SE’s actual performance, since no benchmark data exists for it, but its specification sheet points to a substantially larger GPU with proportionally more execution units and memory throughput.