Intel Arc B570 vs Intel Arc B580 Comparison

Head-to-Head Benchmarks

The benchmark data is unambiguous: the Intel Arc B580 wins every single recorded test against the Intel Arc B570. Across all ten head-to-head comparisons, the B580 takes the victory, with margins ranging from a modest 5.6% to a substantial 16.9%. This is not a close contest; it is a consistent, across-the-board performance advantage.

The largest gap appears in the Passmark DirectX 10 test, where the B580 scores 76 against the B570's 65, a 16.9% lead. This is followed closely by the 3DMark Steel Nomad DX12 test, where the B580 posts 3068 points versus 2649, a 15.8% advantage. These two tests show the B580 at its most dominant, suggesting that in both older DirectX 10 workloads and modern DX12 scenarios, the higher-tier card pulls away decisively.

The Geekbench Vulkan result also shows a significant separation: 109672 for the B580 versus 96844 for the B570, a 13.2% delta. The OpenCL test follows a similar pattern, with the B580 at 92821 and the B570 at 83514, an 11.1% difference. Compute-oriented workloads clearly favor the larger card, though the margin is slightly smaller than in the graphics-focused tests.

The Passmark DirectX 9 test shows an 11.6% gap, with scores of 183 and 164 respectively. The Passmark G3D score, a broad measure of 3D performance, lands at 15748 for the B580 against 14195 for the B570, a 10.9% difference. The Passmark DirectX 11 test narrows the gap to 8.5%, with scores of 128 and 118. The 2D performance test shows a 7.3% gap, with 709 versus 661. The Passmark GPU compute test records the smallest margin at 6.2%, with 7729 against 7281. The Passmark DirectX 12 test is similarly close at 5.6%, with 76 versus 72.

The average benchmark score tells the same story. The B580 averages 23021 across its benchmark suite, while the B570 averages 20556. That is a difference of roughly 2465 points, or about 12% in the B580's favor. The percentile rankings reinforce this: the B580 sits at the 68th percentile among all GPUs, while the B570 sits at the 65th percentile. Both are respectable mid-pack positions, but the B580 is clearly the stronger performer of the two.

Looking at the nearest rivals in the database adds context. The B580's average score of 23021 places it within 0.2% of the AMD Radeon RX 580 2048SP (23061) and 0.6% ahead of the NVIDIA GeForce RTX 2080 (22895). It trails the NVIDIA GeForce RTX 3080 (23172) by 0.7% and the NVIDIA P106-100 (23249) by 1%. The B570, by contrast, sits at 20556, which is 0.1% ahead of the NVIDIA GeForce RTX 3070 Mobile (20534) and 0.4% ahead of the NVIDIA Quadro M4000M (20480). It trails the Intel Arc A750 (20582) by 0.1% and the AMD Radeon R9 M390X (20662) by 0.5%. In short, the B580 competes with a higher tier of hardware than the B570.

The Verdict

The data points to a straightforward conclusion: the Intel Arc B580 is the superior card in every measured category. With a 10-0 record in head-to-head tests, a higher average score, and a higher percentile ranking, there is no benchmark-based argument for choosing the B570 on performance grounds. The B580 leads by double digits in several key tests, including 3DMark Steel Nomad (15.8%), Geekbench Vulkan (13.2%), and Passmark DirectX 10 (16.9%). Even in its weakest showing, the DirectX 12 test, it still manages a 5.6% lead.

Which card should a buyer pick? Strictly from the data, the B580 is the pick for anyone who prioritizes raw performance. Its 12 GB of memory, 456.0 GB/s of bandwidth, and 13.67 TFLOPS of FP32 compute are all clear advantages over the B570's 10 GB, 380.0 GB/s, and 11.52 TFLOPS. The B580 also has more shading units (2560 versus 2304), more texture mapping units (160 versus 144), and more ray tracing cores (20 versus 18). Every architectural advantage translates into a measurable benchmark win.

The B570 is not without merit. It is a fully active product with the same architecture, the same node, and the same feature set. Its 150 W TDP is lower than the B580's 190 W, which could matter in compact builds or systems with tighter power budgets. It also retains the same PCIe 4.0 x8 interface, the same display outputs (1x HDMI 2.1a, 3x DisplayPort 2.1), and the same API support including DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4. For a user who does not need the extra performance headroom, the B570 remains a functional, modern GPU. But the benchmark record is clear: the B580 is the faster card, and the data does not suggest any scenario where the B570 outperforms it.

FAQ

Q: Which GPU wins more benchmark tests?

A: The Intel Arc B580 wins all ten head-to-head tests against the Intel Arc B570. The B570 does not win a single recorded comparison.

Q: What is the biggest performance gap between the two cards?

A: The largest margin is in the Passmark DirectX 10 test, where the B580 leads by 16.9% (76 versus 65). The second largest is in 3DMark Steel Nomad DX12, at 15.8% (3068 versus 2649).

Q: How do the average benchmark scores compare?

A: The B580 has an average benchmark score of 23021, while the B570 averages 20556. The B580 also ranks at the 68th percentile among all GPUs, compared to the B570's 65th percentile.

Q: How does each card compare to its nearest rivals?

A: The B580 is within 0.2% of the AMD Radeon RX 580 2048SP, 0.6% ahead of the NVIDIA GeForce RTX 2080, 0.7% behind the NVIDIA GeForce RTX 3080, and 1% behind the NVIDIA P106-100. The B570 is 0.1% ahead of the NVIDIA GeForce RTX 3070 Mobile, 0.4% ahead of the NVIDIA Quadro M4000M, 0.1% behind the Intel Arc A750, and 0.5% behind the AMD Radeon R9 M390X.

Q: What is the smallest performance difference between the two cards?

A: The smallest margin is in the Passmark DirectX 12 test, where the B580 leads by 5.6% (76 versus 72). The Passmark GPU compute test is close behind at 6.2% (7729 versus 7281).

Q: Do both cards support the same APIs?

A: Yes. Both the B580 and the B570 support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. Neither card has a feature advantage in API support.

Specification Differences

The two cards share the same base chip, BMG-G21, and the same Xe2-HPG architecture, but their specifications diverge in several key areas.

Clock speeds differ: the B580 has a base clock of 2670 MHz and a boost clock of 2670 MHz, while the B570 runs at 2500 MHz for both base and boost. That is a 170 MHz advantage for the B580 on both counts.

Memory is another major differentiator. The B580 comes with 12 GB of GDDR6 memory on a 192-bit bus, delivering 456.0 GB/s of bandwidth. The B570 has 10 GB of GDDR6 on a 160-bit bus, delivering 380.0 GB/s. The B580 also has a higher memory clock at 2375 MHz (19 Gbps effective), which is the same figure listed for the B570, so the bandwidth difference comes from the wider bus and the larger memory pool.

The compute resources also favor the B580. It has 2560 shading units, 160 TMUs, 80 ROPs, and 20 ray tracing cores. The B570 has 2304 shading units, 144 TMUs, 80 ROPs, and 18 ray tracing cores. The ROP count is identical at 80, but the B580 has more of everything else.

Pixel and texture rates reflect these differences. The B580 achieves 213.6 GPixel/s and 427.2 GTexel/s, while the B570 achieves 200.0 GPixel/s and 360.0 GTexel/s. FP32 compute is 13.67 TFLOPS for the B580 and 11.52 TFLOPS for the B570. FP16 compute is 27.34 TFLOPS (2:1) for the B580 and 23.04 TFLOPS (2:1) for the B570.

Power consumption differs as well. The B580 has a TDP of 190 W, while the B570 is rated at 150 W. Both cards use a single 8-pin power connector and share the same suggested PSU rating of 450 W. The B580 is also slightly thicker: its width is listed at 45 mm, while the B570's width is not recorded. Length (272 mm) and height (115 mm) are identical.

The B580 has a launch MSRP of 249 USD and was released on 2024-12-12. The B570 has a launch MSRP of 219 USD and was released on 2025-01-15.

Architecture Differences

Both cards are built on the same Xe2-HPG architecture, Intel's Battlemage generation, and both use the BMG-G21 chip. They are fabricated on the same 5 nm process at TSMC, with the same transistor count of 19,600 million and the same die size of 272 mm². The transistor density is identical at 72.1M per mm².

The architectural differences are therefore ones of scale, not of kind. The B580 enables more of the chip's resources: 2560 shading units versus 2304, 160 TMUs versus 144, and 20 ray tracing cores versus 18. The ROP count is the same at 80. This means the B580 is not a different design; it is a more fully enabled version of the same silicon.

The memory subsystem also differs. The B580 uses a 192-bit bus, while the B570 uses a 160-bit bus. Both use GDDR6 memory at 2375 MHz (19 Gbps effective), but the wider bus on the B580 yields significantly more bandwidth: 456.0 GB/s versus 380.0 GB/s. The B580 also has more memory capacity at 12 GB versus 10 GB.

Feature support is identical. Both cards support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. Both use a PCIe 4.0 x8 interface. Both have the same display outputs: 1x HDMI 2.1a and 3x DisplayPort 2.1. Both are dual-slot cards with a single 8-pin power connector and a suggested PSU of 450 W.

The B570's lower TDP of 150 W, compared to the B580's 190 W, suggests that the B570 is a more power-efficient implementation of the same architecture, likely achieved through lower clocks and fewer active units. The B580's higher clocks (2670 MHz versus 2500 MHz) and additional compute resources explain its higher power draw and its consistent benchmark advantage.

In essence, the architecture is the same; the execution differs. The B580 is the higher-binned, more fully enabled variant, and the benchmark data reflects that at every turn. The B570 is a cut-down version of the same design, trading performance for a lower TDP and a lower launch MSRP.