Intel Arc B390 vs NVIDIA B300 SXM6 AC Comparison

Head-to-Head Benchmarks

The recorded data places these two accelerators in entirely different tiers of the database. The NVIDIA B300 SXM6 AC holds a 100th percentile position among all GPUs, with an average benchmark score of 369831. The Intel Arc B390 sits at the 9th percentile, with an average score of 1482. That raw gap of roughly 250 times is the starting point for any comparison, but the analysis becomes more useful when examining where each device stands against its own nearest rivals.

The B300 SXM6 AC delivers its benchmark score in the Geekbench OpenCL test. Its closest competitor in the database is the NVIDIA B200, which averages 345482. That places the B300 7% ahead of its direct predecessor-class sibling. The margin over the NVIDIA H200 NVL is 10.4%, with the H200 scoring 334891. The AMD Instinct MI300X trails by 16.3%, posting 317994. The NVIDIA L40S, a lower-tier data center card, sits 25% behind with a score of 295763. These deltas show a consistent, if not overwhelming, performance leadership within the server segment. The B300 does not merely edge out one rival; it extends its lead progressively across each lower-ranked competitor.

The Intel Arc B390, by contrast, competes in a much lower performance stratum. Its 1482 score in 3DMark Steel Nomad DX12 places it just ahead of the NVIDIA GeForce GT 520MX, which scores 1463, a 1.3% delta. The GeForce 800M follows at 1460, 1.5% behind. The GeForce GT 625 OEM scores 1446, 2.5% lower. The GeForce GT 710 rounds out the nearest rivals at 1443, 2.7% behind. These are small margins, indicating that the B390 sits at the very edge of its performance class. A 1.3% lead over the GT 520MX is not a commanding position; it is a marginal one within a narrow band of low-end results.

The benchmark data includes no direct head-to-head test between the B390 and the B300 SXM6 AC. The wins count is zero for each device in that category. The comparison therefore relies on their respective database percentiles and nearest-rival deltas. The B300's 100th percentile status means it outperforms virtually every other recorded GPU. The B390's 9th percentile places it below the vast majority of recorded GPUs. The data shows two products designed for different purposes, with no overlap in their competitive fields.

The Verdict

The data indicates that the NVIDIA B300 SXM6 AC is a server-class compute accelerator with a benchmark score that places it at the top of the database. Its nearest rivals are all large-scale data center processors, and it leads each of them by margins ranging from 7% to 25%. The B300 is the clear choice for workloads that require maximum OpenCL compute throughput, as evidenced by its 369831 score and its position above the B200, H200 NVL, MI300X, and L40S.

The Intel Arc B390 is an integrated graphics solution with a 1482 score, placing it in the bottom percentile tier. Its rivals are legacy low-end discrete GPUs, and its leads over them are narrow, between 1.3% and 2.7%. The B390 is appropriate for systems where a dedicated GPU is absent and light graphics acceleration is needed. It does not compete with the B300 in any meaningful sense; the two devices occupy separate segments of the database with no benchmark overlap.

The production status for both products is listed as Active. The B300 SXM6 AC has a release date of September 2025, while the B390 has a release date of January 2026. The B300's predecessor is listed as Server Hopper, and its successor as Server Rubin. The B390 lists no predecessor or successor in the database.

Architecture Differences

The Intel Arc B390 uses the Panther Lake chip with the Xe3-LPG architecture, part of the Arc Graphics-M (Panther Lake) generation. It is built on a 3 nm process at Intel's foundry. The NVIDIA B300 SXM6 AC uses the GB110 chip with the Blackwell Ultra architecture, part of the Server Blackwell (Bxx) generation. It is built on a 5 nm process at TSMC. The transistor counts differ drastically. The B300 packs 208,000 million transistors on a 1628 mm² die, with a transistor density of 127.8M per mm². The B390's transistor count and die size are listed as unknown in the database.

The B300 has 18944 shading units, 592 TMUs, and 24 ROPs. It also includes 592 tensor cores. The B390 has 1536 shading units, 48 TMUs, and 24 ROPs, along with 12 ray tracing cores. The B300 does not list ray tracing cores, and the B390 does not list tensor cores. The memory subsystems are fundamentally different. The B300 uses 288 GB of HBM3e on an 8192-bit bus, delivering 8.19 TB/s of bandwidth. The B390 uses system-shared memory, with system-dependent bandwidth. The B300's memory clock is listed as 2000 MHz with 8 Gbps effective. The B390's memory clock is likewise system-shared.

Clock speeds show the B300 running at a base of 1665 MHz and a boost of 2032 MHz. The B390 runs at a base of 300 MHz and a boost of 2500 MHz. The pixel rates are 48.77 GPixel/s for the B300 and 60.00 GPixel/s for the B390. The texture rates are 1,202.9 GTexel/s for the B300 and 120.0 GTexel/s for the B390. The FP32 throughput is 76.99 TFLOPS for the B300 and 7.680 TFLOPS for the B390. The FP16 rates are 76.99 TFLOPS (1:1) for the B300 and 15.36 TFLOPS (2:1) for the B390.

The B300 is a 1100 W SXM module with a suggested PSU of 1500 W. The B390 is an 80 W IGP with no power connectors. The B300 uses a PCIe 6.0 x16 bus interface and has no display outputs. The B390 uses an IGP bus interface with display outputs listed as portable device dependent. The B300 supports no DirectX, OpenGL, or Vulkan APIs. The B390 supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

FAQ

Q: How does the NVIDIA B300 SXM6 AC compare to its closest rival, the NVIDIA B200?

A: The B300 scores 369831 in Geekbench OpenCL, which is 7% ahead of the B200's 345482.

Q: What is the Intel Arc B390's position relative to the NVIDIA GeForce GT 520MX?

A: The B390 scores 1482 in 3DMark Steel Nomad DX12, which is 1.3% ahead of the GT 520MX's 1463.

Q: Which device has the higher boost clock?

A: The Intel Arc B390 has a boost clock of 2500 MHz, while the NVIDIA B300 SXM6 AC has a boost clock of 2032 MHz.

Q: What memory configurations do the two devices use?

A: The NVIDIA B300 SXM6 AC uses 288 GB of HBM3e on an 8192-bit bus with 8.19 TB/s bandwidth. The Intel Arc B390 uses system-shared memory with system-dependent bandwidth.

Q: What is the transistor count for the NVIDIA B300 SXM6 AC?

A: The B300 has 208,000 million transistors on a 1628 mm² die, built on a 5 nm TSMC process.

Q: Does the Intel Arc B390 support modern graphics APIs?

A: Yes, the B390 supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The B300 lists N/A for all three APIs.

Where Each One Wins

The NVIDIA B300 SXM6 AC wins decisively in raw compute performance. Its 369831 Geekbench OpenCL score is 7% above the B200, 10.4% above the H200 NVL, 16.3% above the MI300X, and 25% above the L40S. The FP32 throughput of 76.99 TFLOPS and FP16 throughput of 76.99 TFLOPS (1:1) indicate balanced compute capability. The 288 GB HBM3e memory with 8.19 TB/s bandwidth provides the capacity and speed required for large model workloads. The 592 tensor cores give it a dedicated path for matrix operations. The 1100 W power envelope and 1500 W suggested PSU confirm its intended deployment in server racks with substantial power delivery.

The Intel Arc B390 wins in the categories that matter for integrated graphics. Its 300 MHz base clock and 2500 MHz boost clock produce a pixel rate of 60.00 GPixel/s, which is higher than the B300's 48.77 GPixel/s. Its 80 W TDP with no power connectors allows it to operate within a host processor's power budget. The 12 ray tracing cores provide hardware acceleration for DirectX 12 Ultimate workloads, a feature the B300 lacks entirely. The B390's support for OpenGL 4.6 and Vulkan 1.4 makes it usable in a wider range of consumer and portable device contexts, whereas the B300 has no display outputs and no graphics API support.

The database records show the B300 as the top percentile GPU, while the B390 sits at the bottom. For server deployments requiring maximum OpenCL throughput, the B300 is the data-backed choice. For portable devices needing integrated graphics with modern API support and low power consumption, the B390 is the only option among these two, as the B300 cannot drive a display. The two products do not overlap in any benchmark category, and the data does not suggest any scenario where they would be cross-shopped.