Intel Arc Pro B390 vs NVIDIA GeForce RTX 5070 Comparison

Where Each One Wins

The Intel Arc Pro B390 and NVIDIA GeForce RTX 5070 occupy entirely different performance strata, and the benchmark data reflects that separation clearly. The RTX 5070 holds an 82nd percentile ranking among all GPUs in the database, while the Arc Pro B390 sits at the 50th percentile. That gap alone frames the competitive picture: the NVIDIA part is a dedicated discrete graphics card aimed at high-throughput workloads, while the Intel part is an integrated graphics processor (IGP) built into the Panther Lake chip.

The RTX 5070 wins in every measurable compute and graphics category where both parts have recorded data. Its PassMark G3D score of 29,137 dwarfs what an integrated solution can deliver. The Arc Pro B390 has no recorded benchmark entries in the database, which means its wins are structural rather than numerical: it consumes 80 W versus 250 W, uses no power connectors, and requires no separate PSU recommendation. For systems where power delivery and physical space are constrained, the Intel part wins by virtue of being an IGP.

The use-case split is stark. The RTX 5070 targets gaming, ray tracing, and GPU compute workloads that demand sustained throughput. Its 30.87 TFLOPS FP32 performance, 48 RT cores, and 192 tensor cores place it in a class where the Arc Pro B390 cannot compete. The Intel part, with 7.680 TFLOPS FP32 and 12 RT cores, serves integrated graphics roles: portable devices, compact systems, and workloads where the GPU shares system memory and bandwidth. The RTX 5070 wins on raw performance, the Arc Pro B390 wins on integration efficiency and power draw.

The Verdict

The data supports a clear verdict for each product. The NVIDIA GeForce RTX 5070 is the choice for any workload where absolute graphics performance matters. Its average benchmark score of 40,377 places it 0.1% ahead of the AMD Radeon Pro 580, 0.8% ahead of the AMD Radeon Pro WX 7100, and 2% ahead of the NVIDIA RTX A500 Mobile. It trails the AMD Radeon Pro 5300 by 1.2%. These are tight margins among discrete workstation and mobile GPUs, but the RTX 5070's 82nd percentile ranking confirms it sits well above the median GPU in the database.

The Intel Arc Pro B390 has no benchmark scores recorded, so its verdict rests on its specifications and role. It is an IGP with system-shared memory, a 300 MHz base clock, and a 2500 MHz boost clock. It delivers 60.00 GPixel/s pixel rate and 120.0 GTexel/s texture rate, figures that are appropriate for integrated graphics but far below the RTX 5070's 201.0 GPixel/s and 482.3 GTexel/s. For users who need a GPU without adding a discrete card, the Arc Pro B390 is the functional option. For users who need performance, the RTX 5070 is the only option in this comparison.

Head-to-Head Benchmarks

The RTX 5070's benchmark results provide the only numerical head-to-head data in this comparison. Its 3DMark Steel Nomad DX12 score of 5,077 demonstrates strong modern API performance. Geekbench results show 172,660 in OpenCL and 178,923 in Vulkan, indicating balanced compute and graphics throughput across both major APIs.

PassMark results break down by DirectX version. The RTX 5070 scores 320 in DirectX 9, 277 in DirectX 11, 180 in DirectX 10, and 108 in DirectX 12. These figures show that older DirectX 9 workloads run relatively faster than newer DirectX 12 workloads in this specific test, though the PassMark G3D score of 29,137 and GPU compute score of 15,787 provide the broader picture. The RTX 5070's average benchmark score of 40,377 aggregates these results into a single comparison point.

The Arc Pro B390 has no recorded benchmark scores, so direct numerical comparison is impossible. Instead, the comparison rests on architectural and specification differences. The RTX 5070 has 6,144 shading units versus 1,536 for the Arc Pro B390, a 4:1 ratio. TMUs stand at 192 versus 48, ROPs at 80 versus 24, and RT cores at 48 versus 12. The RTX 5070 also has 192 tensor cores, which the Intel part lacks entirely. FP32 throughput is 30.87 TFLOPS versus 7.680 TFLOPS, a 4x advantage. FP16 performance is 30.87 TFLOPS on NVIDIA versus 15.36 TFLOPS on Intel, though the Intel figure uses a 2:1 ratio while NVIDIA runs 1:1.

Memory separates the two even further. The RTX 5070 has 12 GB of GDDR7 on a 192-bit bus with 672.0 GB/s bandwidth. The Arc Pro B390 uses system-shared memory with system-dependent bandwidth. That means the Intel part's memory performance depends entirely on the host system's RAM configuration, while the NVIDIA card has dedicated high-speed VRAM.

FAQ

Q: Which GPU has higher raw FP32 performance?

A: The NVIDIA GeForce RTX 5070 delivers 30.87 TFLOPS FP32, while the Intel Arc Pro B390 delivers 7.680 TFLOPS. NVIDIA leads by roughly 4x.

Q: How much memory does each GPU have?

A: The RTX 5070 has 12 GB of GDDR7 memory on a 192-bit bus with 672.0 GB/s bandwidth. The Arc Pro B390 uses system-shared memory with system-dependent bandwidth and no dedicated VRAM.

Q: What is the power consumption difference?

A: The RTX 5070 has a 250 W TDP and requires a 600 W power supply recommendation plus a 16-pin power connector. The Arc Pro B390 has an 80 W TDP, uses no power connectors, and is an integrated graphics processor.

Q: Does the Arc Pro B390 support ray tracing?

A: Yes, it includes 12 RT cores. The RTX 5070 includes 48 RT cores, giving NVIDIA a 4x advantage in ray tracing hardware.

Q: What are the PassMark G3D scores for the RTX 5070?

A: The RTX 5070 scores 29,137 in PassMark G3D and 15,787 in PassMark GPU compute. The Arc Pro B390 has no recorded benchmark scores in the database.

Q: How does the RTX 5070 compare to its nearest rivals?

A: Its average benchmark score of 40,377 puts it 0.1% ahead of the AMD Radeon Pro 580, 0.8% ahead of the AMD Radeon Pro WX 7100, and 2% ahead of the NVIDIA RTX A500 Mobile. It trails the AMD Radeon Pro 5300 by 1.2%.

Architecture Differences

The two GPUs come from different manufacturers, foundries, and process nodes. Intel builds the Arc Pro B390 on a 3 nm process at Intel's own foundry, using the Panther Lake chip with Xe3-LPG architecture. NVIDIA builds the RTX 5070 on a 5 nm process at TSMC, using the GB205 chip with Blackwell 2.0 architecture. The RTX 5070 has 31,100 million transistors on a 263 mm² die, with a transistor density of 118.3M per mm². The Arc Pro B390's transistor count and die size are unknown.

Clock speeds differ significantly. The Arc Pro B390 runs at a 300 MHz base clock and 2500 MHz boost. The RTX 5070 runs at 2325 MHz base and 2512 MHz boost. Memory clocks are 1750 MHz with 28 Gbps effective for NVIDIA, while Intel's memory clock is system-shared and therefore variable.

The RTX 5070 is a dual-slot card measuring 245 mm by 115 mm by 40 mm, using a PCIe 5.0 x16 interface and a single 16-pin power connector. It offers 1x HDMI 2.1b and 3x DisplayPort 2.1b outputs. The Arc Pro B390 is an IGP with no slot width, no power connectors, and portable-device-dependent display outputs. Its bus interface is also IGP, meaning it connects through the processor rather than a PCIe slot.

Both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The RTX 5070 uses 192 tensor cores for AI workloads; the Arc Pro B390 has none. Memory type and bandwidth favor NVIDIA heavily: GDDR7 with 672.0 GB/s versus system-shared with system-dependent bandwidth. The RTX 5070 launched in March 2025 with a 549 USD launch MSRP. The Arc Pro B390 launched in January 2026 with no recorded launch MSRP. The RTX 5070's predecessor is the GeForce 40 series and its successor is the GeForce 60 series. The Arc Pro B390's predecessor is the HD Graphics-WM, and it has no recorded successor.