NVIDIA GeForce RTX 4090 vs NVIDIA GeForce RTX 5090 Mobile Comparison

Head-to-Head Benchmarks

The recorded data presents a sweeping victory for the desktop NVIDIA GeForce RTX 4090 across every single benchmark in the comparison suite. Out of ten head-to-head tests, the RTX 4090 wins all ten, with the RTX 5090 Mobile failing to claim a single test. The margins, however, are far from uniform, and the distribution of those margins tells a compelling story about where each architecture excels and where the mobile part manages to stay within striking distance.

The most dramatic separation appears in the Passmark GPU Compute test. Here, the RTX 4090 scores 26,613 against the RTX 5090 Mobile’s 13,401, a gap of 98.6 percent. That is nearly double the compute output, and it signals a fundamental difference in raw mathematical throughput. The desktop card’s FP32 rating of 82.58 TFLOPS versus the mobile part’s 31.80 TFLOPS, a 2.6x advantage, aligns with this outcome. The compute test is not a marginal win; it is a category demolition.

The 3DMark Steel Nomad DX12 test, which represents modern DirectX 12 gaming workloads, also favors the RTX 4090 heavily. The desktop card posts 9,223 points against 5,871 for the mobile GPU, a 57.1 percent advantage. This is the second-largest delta in the suite and indicates that in current-generation rasterized gaming, the RTX 4090 holds a commanding lead. The RTX 5090 Mobile’s lower pixel rate of 169.7 GPixel/s versus 443.5 GPixel/s, and its texture rate of 496.9 GTexel/s versus 1,290.2 GTexel/s, explain much of this gap.

Geekbench Vulkan shows a 36.9 percent lead for the RTX 4090, with scores of 271,631 against 198,405. Vulkan tends to stress driver efficiency and raw hardware throughput, and here the desktop card’s larger shading unit count, 16,384 versus 10,496, and its 512 tensor cores versus 328, provide a substantial foundation. The OpenCL test is closer in relative terms, with the RTX 4090 at 255,416 and the RTX 5090 Mobile at 201,834, a 26.5 percent margin. OpenCL is often more sensitive to memory bandwidth and compute scheduling, and while the desktop card still wins, the mobile part’s 896.0 GB/s of GDDR7 bandwidth keeps the gap from becoming extreme.

Passmark G3D, a broad gaming and graphics score, gives the RTX 4090 a 27.2 percent win, 38,194 versus 30,034. This test aggregates many DirectX paths and is a useful real-world proxy. The margin here is smaller than in Steel Nomad, which suggests that the mobile GPU’s newer architecture, Blackwell 2.0, handles certain legacy or mixed workloads more efficiently relative to its raw specs. The desktop card’s advantage in shading units, texture mapping units, and render output units is real, but the mobile part is not being blown off the stage in every scenario.

The older DirectX API tests show narrower spreads. Passmark DirectX 12 gives the RTX 4090 only an 8.7 percent lead, 150 versus 138. This is the smallest delta in the entire suite. DirectX 12 is a low-level API that benefits from strong driver optimization and hardware scheduling, and the RTX 5090 Mobile’s Blackwell architecture appears to close the gap significantly here. DirectX 11 shows a 21.2 percent margin, 326 versus 269, while DirectX 10 lands at 22.4 percent, 224 versus 183. DirectX 9, the oldest API tested, gives the RTX 4090 a 22.5 percent win, 397 versus 324. The fact that the newest API produces the smallest gap is notable, as it suggests architectural efficiency in the mobile part that partially compensates for its smaller physical footprint.

Passmark G2D, which measures 2D graphics and desktop composition performance, shows a 22.9 percent lead for the RTX 4090, 1,299 versus 1,057. This is an odd test for gaming GPUs, but it does reflect memory latency and driver overhead. The desktop card’s PCIe 4.0 x16 interface versus the mobile part’s PCIe 5.0 x16 does not show a clear benefit in this metric, as the desktop card still wins comfortably.

Where Each One Wins

The RTX 4090 wins everywhere in raw numbers, but the shape of those wins matters. The dominant victories, compute at 98.6 percent and Steel Nomad at 57.1 percent, point to workloads that saturate the entire GPU: rendering, simulation, AI inference, and high-end gaming at extreme settings. The RTX 4090’s 24 GB of GDDR6X on a 384-bit bus delivers 1.01 TB/s of bandwidth, and its 1,290.2 GTexel/s texture rate means it can feed its 16,384 shading units without bottlenecks. For any task that can scale across massive parallel resources, this card is in a different class.

The RTX 5090 Mobile, despite losing every test, shows its best relative performance in DirectX 12, where it trails by only 8.7 percent, and in OpenCL, where it trails by 26.5 percent. DirectX 12 is the modern gaming API, and an 8.7 percent gap in a laptop GPU against the flagship desktop card of the previous generation is a strong result. The mobile part’s GDDR7 memory, 28 Gbps effective, and its 896.0 GB/s bandwidth, despite a narrower 256-bit bus, keep it competitive in memory-sensitive scenarios. Its lower power envelope, 95 W versus 450 W, means it achieves a significant portion of desktop performance at a fraction of the power draw, though the database does not include thermal or sustained-load measurements.

The RTX 5090 Mobile also performs respectably in legacy DirectX 9 and DirectX 11 tests, with margins around 21 to 22 percent. These APIs are often CPU-bound or driver-bound, and the mobile part’s newer architecture, Blackwell 2.0, appears to handle them efficiently. The 3DMark Steel Nomad test, which is the most demanding modern workload in the suite, is where the mobile part loses the most ground outside of compute. This indicates that the RTX 5090 Mobile is better suited to mainstream gaming and mixed workloads than to extreme high-resolution or compute-heavy tasks.

The Verdict

The data is unambiguous: the desktop NVIDIA GeForce RTX 4090 is the faster GPU in every measured scenario. Its average benchmark score of 60,347 places it in the 88th percentile of all GPUs, while the RTX 5090 Mobile’s average of 45,152 sits in the 84th percentile. The desktop card’s nearest rivals in the database, such as the Intel Arc Pro A60 at 60,326 and the AMD Radeon Pro W6600M at 61,896, are within a few percentage points, which shows that the RTX 4090 is firmly in the upper tier of desktop hardware. The RTX 5090 Mobile, by contrast, sits near the NVIDIA GeForce RTX 4070 Ti, which scores 44,795 and is only 0.8 percent behind, and the Intel Arc A730M at 45,592, which is 1 percent ahead. This places the mobile GPU in a mid-to-high tier, not a flagship tier.

For a user who needs maximum compute throughput, the RTX 4090 is the clear choice. Its 98.6 percent lead in Passmark GPU Compute and its 57.1 percent lead in Steel Nomad make it the only option for heavy rendering, AI workloads, or high-refresh-rate 4K gaming. The launch MSRP of 1,599 USD reflects its desktop flagship status, and its 450 W TDP, triple-slot cooler, and 304 mm length are the physical costs of that performance.

For a user who needs a laptop GPU that can approach desktop performance in modern APIs, the RTX 5090 Mobile is a legitimate option, but the data does not support calling it a direct competitor. Its 8.7 percent gap in DirectX 12 is impressive given its 95 W TDP and IGP form factor, but the 36.9 percent gap in Vulkan and the 27.2 percent gap in G3D show that it is not in the same class for sustained, demanding workloads. The RTX 5090 Mobile’s percentile rank of 84, versus 88 for the RTX 4090, confirms that it is a strong mobile part but not a desktop-class part.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The NVIDIA GeForce RTX 4090 has an average benchmark score of 60,347, while the RTX 5090 Mobile averages 45,152. The RTX 4090 also ranks in the 88th percentile of all GPUs, compared to the 84th percentile for the RTX 5090 Mobile.

Q: How large is the gap in compute performance?

A: In the Passmark GPU Compute test, the RTX 4090 scores 26,613 against 13,401 for the RTX 5090 Mobile, a 98.6 percent advantage. The RTX 4090 also has an FP32 rating of 82.58 TFLOPS versus 31.80 TFLOPS for the mobile part.

Q: Is the RTX 5090 Mobile competitive in any benchmark?

A: The RTX 5090 Mobile comes closest in Passmark DirectX 12, where it trails by only 8.7 percent, scoring 138 against 150. Its OpenCL result of 201,834 is 26.5 percent behind the RTX 4090, and its G3D score of 30,034 is 27.2 percent behind.

Q: What are the memory configurations of the two GPUs?

A: The RTX 4090 has 24 GB of GDDR6X on a 384-bit bus with 1.01 TB/s bandwidth. The RTX 5090 Mobile also has 24 GB of memory, but uses GDDR7 on a 256-bit bus with 896.0 GB/s bandwidth. Both have 24 GB total memory.

Q: Which GPU has more shading units and tensor cores?

A: The RTX 4090 has 16,384 shading units and 512 tensor cores. The RTX 5090 Mobile has 10,496 shading units and 328 tensor cores. The RTX 4090 also has 512 texture mapping units and 176 render output units, compared to 328 and 112 for the mobile part.

Q: What are the power requirements for each GPU?

A: The RTX 4090 has a TDP of 450 W, requires a 1x 16-pin power connector, and a suggested PSU of 850 W. The RTX 5090 Mobile has a TDP of 95 W, uses no external power connectors, and has no suggested PSU listed.

Architecture Differences

The two GPUs come from different generations and different architectures. The RTX 4090 is built on the Ada Lovelace architecture using the AD102 chip, while the RTX 5090 Mobile uses the Blackwell 2.0 architecture with the GB203 chip. Both are manufactured on a 5 nm process at TSMC, but the similarities end there. The AD102 die measures 609 mm² and contains 76,300 million transistors, giving a transistor density of 125.3 million per square millimeter. The GB203 die is smaller at 378 mm², with 45,600 million transistors and a density of 120.6 million per square millimeter. The RTX 4090 is a much larger, more complex chip, and that physical size translates directly into its performance advantage.

The memory architectures differ fundamentally. The RTX 4090 uses GDDR6X with a 384-bit bus, achieving 1.01 TB/s bandwidth. The RTX 5090 Mobile uses GDDR7 with a 256-bit bus, achieving 896.0 GB/s. The newer GDDR7 standard allows for higher effective speeds, 28 Gbps versus 21 Gbps, but the narrower bus limits total bandwidth. The RTX 4090’s pixel rate of 443.5 GPixel/s and texture rate of 1,290.2 GTexel/s dwarf the mobile part’s 169.7 GPixel/s and 496.9 GTexel/s. Ray tracing resources also differ, with the RTX 4090 carrying 128 RT cores against 82 for the RTX 5090 Mobile.

The RTX 5090 Mobile does bring a newer bus interface, PCIe 5.0 x16, versus the RTX 4090’s PCIe 4.0 x16. However, the database does not include tests that specifically isolate PCIe bandwidth effects, so the practical impact of this difference is not directly measured. Both GPUs support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4, so API compatibility is identical.

Specification Differences

The two GPUs differ across nearly every specification field. The RTX 4090 has a base clock of 2235 MHz and a boost clock of 2520 MHz, while the RTX 5090 Mobile has a base clock of 990 MHz and a boost clock of 1515 MHz. Memory speed differs as well, with the RTX 4090 at 1313 MHz or 21 Gbps effective, versus 1750 MHz or 28 Gbps effective for the mobile part. The RTX 4090’s FP32 compute is 82.58 TFLOPS, more than double the RTX 5090 Mobile’s 31.80 TFLOPS, and the same applies to FP16, which is also 82.58 TFLOPS versus 31.80 TFLOPS.

Physical specifications are completely different. The RTX 4090 is a triple-slot desktop card that is 304 mm long, 137 mm high, and 61 mm wide, with a 1x 16-pin power connector and a 450 W TDP. The RTX 5090 Mobile is an IGP, meaning it is integrated into a laptop motherboard, with no dimensions listed, no power connectors, and a 95 W TDP. The RTX 4090 has a suggested PSU of 850 W, while the mobile part has none. Display outputs also differ: the RTX 4090 offers 1x HDMI 2.1 and 3x DisplayPort 1.4a, while the RTX 5090 Mobile’s outputs are listed as portable device dependent.

The production status differs as well. The RTX 4090 is end-of-life, released on September 19, 2022, with a launch MSRP of 1,599 USD, while the RTX 5090 Mobile is an active product, released on March 26, 2025, with no launch MSRP listed. The RTX 4090’s predecessor is the GeForce 30 series, and its successor is the GeForce 50 series. The RTX 5090 Mobile’s predecessor is the GeForce 40 Mobile, and it has no successor listed. The RTX 4090 belongs to the GeForce 40-series generation, while the RTX 5090 Mobile belongs to the GeForce 50 Mobile generation.