Intel Arc B390 vs Intel Arc B570 Comparison

FAQ

Q: What is the performance difference between the Intel Arc B390 and the Intel Arc B570 in the recorded 3DMark Steel Nomad DX12 test?

A: The Intel Arc B570 scores 2649, while the Intel Arc B390 scores 1482. The B570 wins this test by 44.1%.

Q: How do the two GPUs compare in terms of memory capacity and type?

A: The Intel Arc B570 uses 10 GB of GDDR6 memory with a 160-bit bus and 380.0 GB/s bandwidth. The Intel Arc B390 uses system shared memory, with its bandwidth and bus width listed as system dependent.

Q: What are the thermal design power ratings for these two GPUs?

A: The Intel Arc B390 has an 80 W TDP, while the Intel Arc B570 has a 150 W TDP.

Q: Which GPU uses a smaller manufacturing process node?

A: The Intel Arc B390 is built on a 3 nm process, while the Intel Arc B570 uses a 5 nm process.

Q: What is the average benchmark score for each GPU across all recorded tests?

A: The Intel Arc B570 has an average benchmark score of 20556, while the Intel Arc B390 has an average benchmark score of 1482.

Q: How do the two GPUs rank against all GPUs in the database?

A: The Intel Arc B570 sits at the 65th percentile, while the Intel Arc B390 is at the 9th percentile.

Architecture Differences

The Intel Arc B390 and Intel Arc B570 represent two distinct architectural generations from Intel. The B390 is based on the Panther Lake chip with the Xe3-LPG architecture, belonging to the Arc Graphics-M (Panther Lake) generation. The B570, in contrast, uses the BMG-G21 chip with the Xe2-HPG architecture, part of the Battlemage (Arc 5) generation. This architectural split reflects a mobile-oriented integrated design versus a discrete add-in board design.

The manufacturing process differs substantially. The B390 is fabricated on a 3 nm node at Intel's own foundry, while the B570 uses a 5 nm node from TSMC. The B570 discloses 19,600 million transistors on a 272 mm² die, giving a transistor density of 72.1M per mm². The B390 lists its transistor count and die size as unknown.

The compute resource allocation shows a clear hierarchy. The B390 has 1536 shading units, 48 texture mapping units, 24 raster output pipelines, and 12 ray tracing cores. The B570 scales up to 2304 shading units, 144 TMUs, 80 ROPs, and 18 RT cores. This represents a 50% increase in shading units and a 200% increase in TMUs and ROPs for the B570.

Clock behavior differs notably between the two. The B390 runs at a 300 MHz base clock and boosts to 2500 MHz, a wide frequency range typical of power-managed integrated graphics. The B570 runs at a flat 2500 MHz for both base and boost clocks, indicating a fixed operating point. Memory clocks also diverge: the B390 uses system shared memory, while the B570 has a dedicated 2375 MHz GDDR6 configuration rated at 19 Gbps effective.

The B570's memory subsystem is fully specified: 10 GB GDDR6, 160-bit bus, 380.0 GB/s bandwidth. The B390's memory is entirely system dependent, with size, type, bus width, and bandwidth all listed as system shared or system dependent. This is a fundamental difference in how each GPU accesses data.

Power delivery and physical design differ as well. The B390 is an integrated graphics processor (IGP) with no power connectors and an 80 W TDP. The B570 is a dual-slot discrete card requiring a single 8-pin power connector, with a 150 W TDP and a suggested power supply of 450 W. The B570 measures 272 mm in length and 115 mm in height. The B390 has no listed dimensions.

The B570 offers a richer set of display outputs: 1x HDMI 2.1a and 3x DisplayPort 2.1. The B390's display outputs are listed as portable device dependent. Both GPUs support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

The Verdict

The recorded data places these two GPUs in entirely different performance tiers. The Intel Arc B390's average benchmark score of 1482 places it at the 9th percentile of all GPUs. Its nearest rivals in the database are the NVIDIA GeForce GT 520MX (1463, 1.3% behind), the NVIDIA GeForce 800M (1460, 1.5% behind), the NVIDIA GeForce GT 625 OEM (1446, 2.5% behind), and the NVIDIA GeForce GT 710 (1443, 2.7% behind). This cluster of scores indicates the B390 operates at an entry-level, low-power segment.

The Intel Arc B570's average score of 20556 places it at the 65th percentile. Its nearest rivals are the NVIDIA GeForce RTX 3070 Mobile (20534, 0.1% behind), the Intel Arc A750 (20582, 0.1% ahead), the NVIDIA Quadro M4000M (20480, 0.4% behind), and the AMD Radeon R9 M390X (20662, 0.5% ahead). The B570 sits in a competitive mid-range band, essentially matching the RTX 3070 Mobile and the Arc A750 within a fraction of a percent.

The B390 is an integrated solution with system shared memory and a 3 nm process, designed for portability and low power. Its 80 W TDP, lack of power connectors, and IGP bus interface confirm a mobile or compact platform role. The B570 is a discrete, dual-slot card with dedicated GDDR6 memory, a 150 W TDP, and a PCIe 4.0 x8 interface, targeting desktop builds that need a dedicated mid-range GPU.

For users constrained to an integrated platform, the B390 provides a baseline 3D capability with modern API support. For those building or upgrading a desktop with a PCIe slot and a 450 W power supply, the B570 delivers significantly higher compute throughput, 50% more shading units, and a dedicated memory subsystem with 380.0 GB/s bandwidth. The data does not support any other conclusion: the B570 is the higher-performing part by a wide margin, and the B390 serves a fundamentally different, lower-power integrated niche.

Specification Differences

| Specification | Intel Arc B390 | Intel Arc B570 |

|---|---|---|

| Chip | Panther Lake | BMG-G21 |

| Architecture | Xe3-LPG | Xe2-HPG |

| Generation | Arc Graphics-M (Panther Lake) | Battlemage (Arc 5) |

| Process Node | 3 nm | 5 nm |

| Foundry | Intel | TSMC |

| Transistors | unknown | 19,600 million |

| Die Size | unknown | 272 mm² |

| Transistor Density | null | 72.1M / mm² |

| Base Clock | 300 MHz | 2500 MHz |

| Boost Clock | 2500 MHz | 2500 MHz |

| Memory Clock | System Shared | 2375 MHz 19 Gbps effective |

| Memory Size | System Shared | 10 GB |

| Memory Type | System Shared | GDDR6 |

| Memory Bus Width | System Shared | 160 bit |

| Memory Bandwidth | System Dependent | 380.0 GB/s |

| Shading Units | 1536 | 2304 |

| TMUs | 48 | 144 |

| ROPs | 24 | 80 |

| RT Cores | 12 | 18 |

| Pixel Rate | 60.00 GPixel/s | 200.0 GPixel/s |

| Texture Rate | 120.0 GTexel/s | 360.0 GTexel/s |

| FP32 | 7.680 TFLOPS | 11.52 TFLOPS |

| FP16 | 15.36 TFLOPS (2:1) | 23.04 TFLOPS (2:1) |

| TDP | 80 W | 150 W |

| Slot Width | IGP | Dual-slot |

| Power Connectors | None | 1x 8-pin |

| Suggested PSU | null | 450 W |

| Bus Interface | IGP | PCIe 4.0 x8 |

| Display Outputs | Portable Device Dependent | 1x HDMI 2.1a3x DisplayPort 2.1 |

| Dimensions | null | 272 mm 10.7 inches length, 115 mm 4.5 inches height |

| Release Date | 2026-01-26 | 2025-01-15 |

| Predecessor | null | Alchemist |

| Launch MSRP | null | 219 USD |

Head-to-Head Benchmarks

The database records a single shared benchmark between the two GPUs: 3DMark Steel Nomad DX12. The Intel Arc B570 scores 2649, and the Intel Arc B390 scores 1482. The B570 wins by 44.1%. This is the only direct head-to-head measurement available, and it establishes a decisive performance gap.

The B570's win in Steel Nomad is consistent with its broader specification advantages. Its FP32 throughput of 11.52 TFLOPS is 50% higher than the B390's 7.680 TFLOPS. The pixel rate of 200.0 GPixel/s on the B570 is more than triple the B390's 60.00 GPixel/s. Texture rate follows the same pattern: 360.0 GTexel/s versus 120.0 GTexel/s, a 200% advantage.

The B570 also holds the memory advantage in this test. Its 380.0 GB/s dedicated bandwidth versus the B390's system dependent memory means the discrete card does not compete with the CPU for memory access. The 10 GB GDDR6 allocation with a 160-bit bus provides a stable, high-throughput memory environment.

The B390's only recorded benchmark is the same Steel Nomad test, so its average benchmark score of 1482 equals this single result. The B570 has additional recorded tests including Geekbench OpenCL (83514), Geekbench Vulkan (96844), and multiple Passmark tests, with a G3D score of 14195 and a GPU compute score of 7281. These additional data points raise the B570's average to 20556, but the head-to-head comparison remains anchored to the Steel Nomad result.

The percentile gap reinforces the benchmark outcome. The B570 at the 65th percentile sits far above the B390 at the 9th percentile. The rival clusters tell the same story: the B390 trades results with legacy NVIDIA parts like the GT 710 and GT 520MX, while the B570 matches current mobile and desktop mid-range parts like the RTX 3070 Mobile and Arc A750.

In the recorded head-to-head data, the B570 wins the only contested benchmark, and the wider database confirms the separation in performance class.