Intel Arc B580 vs NVIDIA B300 SXM6 AC Comparison

Head-to-Head Benchmarks

The database contains a single overlapping benchmark between the Intel Arc B580 and the NVIDIA B300 SXM6 AC: Geekbench OpenCL. The NVIDIA B300 SXM6 AC delivers a score of 369,831, while the Intel Arc B580 records 92,821. This represents a delta of -74.9% for the Intel part, meaning the NVIDIA accelerator is roughly four times faster in this compute-oriented workload. The margin is substantial and consistent with the positioning of the two products: one is a consumer graphics card, the other a server-grade compute module.

In absolute terms, the NVIDIA B300 SXM6 AC outscores the Intel Arc B580 by 276,010 points in Geekbench OpenCL. This is the only direct head-to-head measurement available, so the comparison rests entirely on this single workload. The Intel Arc B580 does not win any of the recorded benchmark comparisons (winsA: 0), while the NVIDIA B300 SXM6 AC wins the sole contest (winsB: 1). For context, the Intel Arc B580's average benchmark score across all recorded tests is 23,021, placing it at the 68th percentile among all GPUs in the database. The NVIDIA B300 SXM6 AC, by contrast, holds a 100th percentile ranking with an average score of 369,831, which places it at the very top of the distribution.

Relative to its nearest rivals, the Intel Arc B580 sits in a tightly clustered competitive band. Its average score of 23,021 is within 1% of the AMD Radeon RX 580 2048SP (23,061, delta -0.2%), the NVIDIA GeForce RTX 2080 (22,895, delta 0.6%), the NVIDIA GeForce RTX 3080 (23,172, delta -0.7%), and the NVIDIA P106-100 (23,249, delta -1%). This indicates that in aggregate compute performance, the Arc B580 is effectively indistinguishable from those four cards, with deltas small enough to be within measurement noise.

The NVIDIA B300 SXM6 AC, in contrast, leads its nearest rivals by substantial margins. It is 7% ahead of the NVIDIA B200 (345,482), 10.4% ahead of the NVIDIA H200 NVL (334,891), 16.3% ahead of the AMD Instinct MI300X (317,994), and 25% ahead of the NVIDIA L40S (295,763). These are large, consistent gaps that confirm the B300 SXM6 AC sits in a performance tier above even the most capable existing accelerators in the database.

Where Each One Wins

The Intel Arc B580 is a consumer-focused graphics card with a broad suite of recorded benchmarks spanning DirectX 9 through DirectX 12, OpenCL, Vulkan, and 2D/3D graphics tests. Its strongest recorded result is Geekbench Vulkan at 109,672, followed by Geekbench OpenCL at 92,821. In Passmark tests, it scores 183 in DirectX 9, 128 in DirectX 11, 76 in DirectX 10, and 76 in DirectX 12. The Passmark G3D score is 15,748, and GPU compute is 7,729. The 2D score is 709. These results indicate a card that handles legacy and modern graphics APIs, with Vulkan being its strongest compute interface. The Arc B580 also supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4, making it suitable for gaming and general-purpose graphics workloads.

The NVIDIA B300 SXM6 AC has only one recorded benchmark, Geekbench OpenCL at 369,831. It does not support DirectX, OpenGL, or Vulkan (all listed as N/A), and it has no display outputs. This makes it a compute-only accelerator aimed at server and datacenter workloads such as AI training, inference, and scientific simulation. Its 592 tensor cores and 76.99 TFLOPS of FP16 (at a 1:1 ratio) suggest it is optimized for tensor-heavy operations rather than rasterization. The ROP count of 24 is low relative to its shader count of 18,944, further indicating that pixel throughput is not a priority.

In practical terms, the Arc B580 wins in any scenario requiring graphics output, display connectivity, or legacy API support. The B300 SXM6 AC wins decisively in raw compute throughput, as evidenced by the OpenCL result, and its architecture is designed for high-density compute in server environments. The two products do not compete in the same market segment; the Arc B580 targets desktop graphics, while the B300 SXM6 AC targets accelerated computing in scale-out deployments.

FAQ

Q: Which GPU has the higher Geekbench OpenCL score?

A: The NVIDIA B300 SXM6 AC scores 369,831, while the Intel Arc B580 scores 92,821. The NVIDIA part is 74.9% higher, equating to a delta of -74.9% for the Intel card.

Q: How does the Intel Arc B580 compare to its nearest rivals in average benchmark score?

A: The Arc B580's average score is 23,021. It trails the AMD Radeon RX 580 2048SP by 0.2% (23,061), leads the NVIDIA GeForce RTX 2080 by 0.6% (22,895), trails the NVIDIA GeForce RTX 3080 by 0.7% (23,172), and trails the NVIDIA P106-100 by 1% (23,249).

Q: What is the percentile ranking of each GPU?

A: The Intel Arc B580 sits at the 68th percentile among all GPUs in the database. The NVIDIA B300 SXM6 AC sits at the 100th percentile, meaning no other recorded GPU scores higher.

Q: Does the NVIDIA B300 SXM6 AC support DirectX or Vulkan?

A: No. Its API support is listed as N/A for DirectX, OpenGL, and Vulkan. It also has no display outputs, confirming it is a compute-only accelerator.

Q: What memory configurations do the two GPUs use?

A: The Intel Arc B580 uses 12 GB of GDDR6 on a 192-bit bus, delivering 456.0 GB/s of bandwidth. The NVIDIA B300 SXM6 AC uses 288 GB of HBM3e on an 8192-bit bus, delivering 8.19 TB/s of bandwidth.

Q: How does the NVIDIA B300 SXM6 AC compare to its nearest rivals?

A: It leads the NVIDIA B200 by 7% (345,482), the NVIDIA H200 NVL by 10.4% (334,891), the AMD Instinct MI300X by 16.3% (317,994), and the NVIDIA L40S by 25% (295,763).

Specification Differences

The two GPUs differ in nearly every measurable specification. The Intel Arc B580 uses the BMG-G21 chip based on Xe2-HPG architecture, part of the Battlemage (Arc 5) generation, while the NVIDIA B300 SXM6 AC uses the GB110 chip based on Blackwell Ultra architecture, part of the Server Blackwell (Bxx) generation. Both are fabricated on a 5 nm process at TSMC, but the transistor counts diverge sharply: the Arc B580 has 19,600 million transistors on a 272 mm² die, yielding a density of 72.1M / mm². The B300 SXM6 AC has 208,000 million transistors on a 1,628 mm² die, yielding a density of 127.8M / mm².

Clock speeds also differ. The Arc B580 runs at a base and boost of 2670 MHz, with memory at 2375 MHz (19 Gbps effective). The B300 SXM6 AC has a base clock of 1665 MHz and a boost of 2032 MHz, with memory at 2000 MHz (8 Gbps effective). Despite lower clocks, the B300 SXM6 AC's massive memory bus and HBM3e type give it 8.19 TB/s of bandwidth versus 456.0 GB/s for the Arc B580. Memory capacity is 288 GB versus 12 GB.

Compute resources differ substantially. The Arc B580 has 2,560 shading units, 160 TMUs, 80 ROPs, 20 RT cores, and no tensor cores. The B300 SXM6 AC has 18,944 shading units, 592 TMUs, 24 ROPs, no RT cores, and 592 tensor cores. Pixel rates are 213.6 GPixel/s for the Arc B580 and 48.77 GPixel/s for the B300 SXM6 AC. Texture rates are 427.2 GTexel/s and 1,202.9 GTexel/s, respectively. FP32 performance is 13.67 TFLOPS for the Intel card and 76.99 TFLOPS for the NVIDIA card. FP16 is 27.34 TFLOPS (2:1) for the Arc B580 and 76.99 TFLOPS (1:1) for the B300 SXM6 AC.

Power and physical specifications are also divergent. The Arc B580 has a TDP of 190 W, a suggested PSU of 450 W, a dual-slot form factor, and a single 8-pin power connector. The B300 SXM6 AC has a TDP of 1100 W, a suggested PSU of 1500 W, an SXM Module form factor, and no power connectors listed. The Arc B580 uses PCIe 4.0 x8 and has 1x HDMI 2.1a and 3x DisplayPort 2.1 outputs. The B300 SXM6 AC uses PCIe 6.0 x16 and has no display outputs. The Arc B580 measures 272 mm in length, 115 mm in height, and 45 mm in width. The B300 SXM6 AC has no recorded dimensions.

Architecture Differences

The Intel Arc B580 is built on the Xe2-HPG architecture, which is designed for consumer graphics with support for DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. It includes dedicated RT cores (20 of them) for hardware-accelerated ray tracing, a feature entirely absent from the B300 SXM6 AC, which lists no RT cores. The Arc B580's FP16 performance is delivered at a 2:1 ratio relative to FP32, indicating a consumer-oriented design where half-precision is a secondary mode.

The NVIDIA B300 SXM6 AC is built on the Blackwell Ultra architecture, which is a server-focused design. It includes 592 tensor cores, which are not present in the Arc B580 at all. FP16 runs at a 1:1 ratio with FP32, meaning the card can sustain full-rate half-precision compute, a common requirement for AI and machine learning workloads. The B300 SXM6 AC has no graphics API support (DirectX, OpenGL, and Vulkan are all N/A), no display outputs, and a much higher transistor density (127.8M / mm² versus 72.1M / mm²), reflecting a design optimized for compute density rather than graphics throughput.

The memory architectures are fundamentally different as well. The Arc B580 uses GDDR6 on a 192-bit bus, a conventional choice for consumer cards. The B300 SXM6 AC uses HBM3e on an 8192-bit bus, which is typical for high-performance accelerators where memory bandwidth is critical. The B300 SXM6 AC's bandwidth of 8.19 TB/s is nearly 18 times that of the Arc B580's 456.0 GB/s. The B300 SXM6 AC also carries a successor (Server Rubin) and a predecessor (Server Hopper), whereas the Arc B580's predecessor is Alchemist and its successor is not listed.

The Intel Arc B580 was released on December 12, 2024, while the NVIDIA B300 SXM6 AC was released on September 10, 2025. Both are currently listed as Active in production. The Arc B580 has a launch MSRP of 249 USD, while the B300 SXM6 AC has no recorded launch MSRP. The Arc B580's power delivery is via a single 8-pin connector with a 450 W suggested PSU, while the B300 SXM6 AC's power requirements are much higher, with a 1100 W TDP and a 1500 W suggested PSU, though no specific connector is listed.