AMD Radeon RX 6700 XT vs Intel Arc B580 Comparison

Where Each One Wins

The benchmark data splits this comparison into two very distinct personalities. The AMD Radeon RX 6700 XT dominates legacy and general-purpose workloads, while the Intel Arc B580 takes the crown in modern DirectX 12 and compute-heavy APIs. The scoreboard reads 7 wins for AMD and 3 for Intel, but the margins tell a more nuanced story.

The AMD Radeon RX 6700 XT sweeps the PassMark suite entirely. Its biggest victory comes in DirectX 10, where it scores 124 against Intel's 76, a 63.2% advantage. DirectX 11 shows a 35.2% lead (173 versus 128), and even DirectX 9, a legacy API, sees AMD ahead by 32.2% (242 vs 183). The DirectX 12 test is nearly a dead heat: AMD scores 77, Intel scores 76, a 1.3% margin that is effectively noise. Beyond gaming APIs, AMD also wins the 2D and 3D PassMark tests by 34.1% (951 vs 709) and 25.6% (19786 vs 15748), respectively. The GPU compute PassMark test also favors AMD: 9336 vs 7729, a 20.8% edge.

The Intel Arc B580's wins are concentrated in the newest and most demanding workloads. In 3DMark Steel Nomad DX12, Intel posts 3068 against AMD's 2435, a 20.6% advantage. The gap is even larger in Geekbench OpenCL: Intel scores 92821 versus AMD's 44152, more than doubling AMD's result with a 52.4% lead. In Geekbench Vulkan, Intel takes 109672 against AMD's 93500, a 14.7% margin.

The pattern is consistent: AMD owns older APIs and rasterization-focused tests, while Intel's Xe2 architecture shines in modern, low-level APIs and parallel compute. For users running current DirectX 12 games, the Intel card is the stronger pick. For everything else, AMD holds the lead.

The Verdict

The data does not point to a single winner; it points to two different buyers.

Pick the AMD Radeon RX 6700 XT if your workload relies on DirectX 11, DirectX 9, or general 3D rendering. The 63.2% lead in DirectX 10 and 35.2% lead in DirectX 11 are decisive. The 25.6% margin in PassMark G3D further cements its position for traditional rasterized gaming. Its overall average benchmark score of 27425 places it at the 72nd percentile of all GPUs, putting it on par with the NVIDIA GeForce RTX 4070 Mobile (27435, 0% delta) and slightly below the RTX 3090 (27565, -0.5%). This is a card that competes with much more expensive hardware in the aggregate.

Pick the Intel Arc B580 if your priorities are modern DirectX 12 titles and compute-heavy applications. The 20.6% win in 3DMark Steel Nomad is not a small margin, and the 52.4% blowout in Geekbench OpenCL indicates a clear compute advantage. However, its overall average of 23021 sits at the 68th percentile, which is notably lower than AMD's 72nd. It sits within 1% of the RTX 2080 (22895, 0.6% delta) and RTX 3080 (23172, -0.7%), meaning it is a completely different performance class than the RX 6700 XT in aggregate terms.

The data says the RX 6700 XT is the faster card overall. The Intel card is faster only in the specific scenarios where its new architecture pays off. If most of your gaming or rendering uses older APIs, the AMD is the clear choice. If you live in the newest titles and compute, the Intel has a real advantage.

Architecture Differences

The two cards come from different eras and design philosophies. The AMD Radeon RX 6700 XT uses the Navi 22 chip on the RDNA 2.0 architecture, a 7 nm design from TSMC. The Intel Arc B580 uses the BMG-G21 chip based on Xe2-HPG, built on a 5 nm process at TSMC. The Intel chip is newer and denser: 19,600 million transistors on a 272 mm² die (72.1M / mm²) versus AMD's 17,200 million transistors on a 335 mm² die (51.3M / mm²). Intel achieves more transistors in a smaller physical package.

Both cards feature the same core counts: 2560 shading units and 160 texture mapping units. They differ in the render output stage. The AMD has 64 ROPs, while the Intel has 80 ROPs. The Intel therefore pushes more pixels: 213.6 GPixel/s vs AMD's 165.2 GPixel/s. The texture rates are closer, with Intel at 427.2 GTexel/s and AMD at 413.0 GTexel/s.

The ray tracing implementations are very different in count. AMD includes 40 ray tracing cores, while Intel has 20. The FP32 performance is marginally higher on Intel: 13.67 TFLOPS vs 13.21 TFLOPS. FP16 performance, at 27.34 TFLOPS vs 26.43 TFLOPS, follows the same pattern.

Memory systems are similar in capacity and bus width: both have 12 GB of GDDR6 across a 192-bit bus. The Intel runs faster memory, achieving 456.0 GB/s of bandwidth against AMD's 384.0 GB/s. That 72 GB/s gap is a meaningful architectural trait for texture-heavy workloads.

Power consumption is also different. The Intel's TDP is 190 W, while AMD's is 230 W. The AMD requires a 6-pin and an 8-pin connector, while the Intel uses a single 8-pin. The suggested PSU is 450 W for the Intel and 550 W for the AMD. The Intel is a more power-efficient card per watt of performance, based on the TDP figures. The AMD is a dual-slot card, as is the Intel, but the AMD is slightly shorter (267 mm vs 272 mm), lower (110 mm vs 115 mm), and thinner (40 mm vs 45 mm). The AMD runs a PCIe 4.0 x16 interface, while the Intel uses PCIe 4.0 x8, which could be a limitation on older platforms with fewer PCIe lanes.

Display outputs differ, too: the AMD has 1x HDMI 2.1 and 3x DisplayPort 1.4a, while the Intel has 1x HDMI 2.1a and 3x DisplayPort 2.1. Intel supports the newer DisplayPort standard, which is relevant for high refresh rate 4K and 8K displays.

The AMD is from the Radeon RX 6000 series, end-of-life, released on the 2nd of March 2021. The Intel is from the Battlemage (Arc 5) generation, released on the 12th of December 2024, and it is still an active product. The AMD's predecessor is Navi, its successor Navi III. The Intel's predecessor is Alchemist, with no successor announced. The AMD's launch MSRP was 479 USD; the Intel's was 249 USD. The production status is the key: one is a legacy product, the other is a current one.

FAQ

Q: Which GPU has higher raw compute performance?

The Intel Arc B580 has higher FP32 performance at 13.67 TFLOPS, compared to the AMD Radeon RX 6700 XT's 13.21 TFLOPS. The Intel also leads in FP16: 27.34 TFLOPS vs 26.43 TFLOPS.

Q: Which GPU is better for modern DirectX 12 gaming?

A: The Intel Arc B580 wins the 3DMark Steel Nomad DX12 test with a score of 3068 against AMD's 2435, a 20.6% advantage. However, in the PassMark DirectX 12 test, the AMD is marginally ahead at 77 vs 76, a 1.3% difference. The Intel is the stronger performer in the heavier DX12 load.

Q: How much faster is the AMD in DirectX 11?

A: The AMD Radeon RX 6700 XT scores 173 in the PassMark DirectX 11 test, while the Intel Arc B580 scores 128. This represents a 35.2% margin in favor of the AMD.

Q: Which GPU has more memory bandwidth?

A: The Intel Arc B580 has a higher bandwidth of 456.0 GB/s, while the AMD Radeon RX 6700 XT has 384.0 GB/s. Both share 12 GB of GDDR6 memory and a 192-bit bus, but the Intel's memory clock is 2375 MHz compared to AMD's 2000 MHz.

Q: Which GPU is more power-hungry?

A: The AMD Radeon RX 6700 XT has a TDP of 230 W and requires a 550 W PSU, while the Intel Arc B580 has a TDP of 190 W and requires a 450 W PSU. The AMD's power connectors are 1x 6-pin + 1x 8-pin, while the Intel uses a single 8-pin.

Q: Which GPU is better for compute workloads?

A: The Intel Arc B580 dominates in OpenCL, scoring 92821 versus AMD's 44152, a 52.4% delta. The Intel also leads in Vulkan (109672 vs 93500, a 14.7% delta). The AMD scores higher in the PassMark GPU compute test (9336 vs 7729, a 20.8% delta). The Intel is better for OpenCL-heavy compute.

Head-to-Head Benchmarks

The most decisive AMD victory is in the legacy DirectX 10 test, where the RX 6700 XT scores 124 versus the Arc B580's 76. That 63.2% gap is the single largest margin in either direction. The DirectX 11 test follows with a 35.2% win (173 to 128), and DirectX 9 shows a 32.2% lead (242 vs 183). The DirectX 12 PassMark test is almost a tie, with AMD leading 1.3% (77 to 76), so the AMD takes all four legacy PassMark API tests.

The PassMark 2D test goes to AMD by 34.1% (951 vs 709). The PassMark 3D test is more lopsided: AMD scores 19786, Intel scores 15748, a 25.6% advantage. The PassMark GPU Compute test adds another win for AMD with 9336 vs 7729, a 20.8% margin.

The Intel's biggest victory comes in Geekbench OpenCL, where it scores 92821 against AMD's 44152. That is a 52.4% delta, and the largest absolute score difference in the entire comparison. The Intel nearly doubles AMD's OpenCL output, which is a strong signal for compute-oriented users.

The 3DMark Steel Nomad DX12 test is another clear Intel advantage, with 3068 points versus AMD's 2435, a 20.6% delta. This is a modern, demanding DX12 benchmark, and the Intel's newer architecture shines here.

Geekbench Vulkan shows Intel winning again, 109672 to AMD's 93500, a 14.7% margin.

The overall pattern is consistent across every benchmark. AMD's wins are concentrated in the PassMark suite, with margins ranging from 20.8% to 63.2%. Intel's wins are fewer but substantial, with margins from 14.7% to 52.4%. The two cards are not close in any specific test, except for the PassMark DirectX 12 tie. Which card you choose depends entirely on which API and workload you care about. The data does not suggest a balanced alternative; it suggests an either/or choice between legacy rasterization and modern compute.