Intel Arc Pro B370 vs Intel Arc Pro B390 Comparison

Head-to-Head Benchmarks

The recorded database contains no direct head-to-head benchmark runs between the Intel Arc Pro B370 and the Intel Arc Pro B390. Both entries list an empty benchmark array, a wins count of zero for each part, and an average benchmark score of zero. This absence of measured data means the comparison must rely entirely on the architectural and specification fields provided in the database.

The only quantitative performance indicators available are the derived throughput figures. The Arc Pro B390 delivers 7.680 TFLOPS of FP32 compute, which is exactly 25% higher than the 6.144 TFLOPS of the Arc Pro B370. The FP16 figures follow the same proportional pattern: 15.36 TFLOPS (2:1) for the B390 versus 12.29 TFLOPS (2:1) for the B370, again a 25% advantage. Pixel fill rates show the B390 at 60.00 GPixel/s against 48.00 GPixel/s for the B370, a 25% lead. Texture fill rates also scale by 25%, with the B390 reaching 120.0 GTexel/s versus 96.00 GTexel/s for the B370.

These consistent 25% deltas across four separate compute and fill-rate metrics indicate a deliberate scaling pattern rather than a rebalanced design. The boost clock difference contributes to this: the B390 boosts to 2500 MHz while the B370 boosts to 2400 MHz, a 4.17% clock advantage. The remaining gap is explained by the larger execution resource count in the B390. Since no measured benchmark scores exist, the database cannot confirm whether real-world applications achieve the theoretical scaling, but the specification sheet strongly implies a linear performance uplift across most workloads.

Architecture Differences

Both graphics processors share the same fundamental design: they are built on the Panther Lake chip, use the Xe3-LPG architecture, and belong to the Arc Graphics-WM (Panther Lake) generation. Both are manufactured on a 3 nm process at Intel. The transistor count and die size are listed as unknown for both, so no comparison is possible on those fronts.

The execution resource counts differ exactly by 20%. The Arc Pro B370 has 1280 shading units, 40 texture mapping units, 20 raster operations pipelines, and 10 ray tracing cores. The Arc Pro B390 scales each category by 20%: 1536 shading units, 48 TMUs, 24 ROPs, and 12 RT cores. This uniform scaling suggests the B390 is a full-width implementation of the same silicon, while the B370 is a partially disabled configuration. The tensor core field is null for both, so neither part exposes dedicated tensor hardware in the database.

Clock behavior is nearly identical: both have a 300 MHz base clock, but the B390 boosts to 2500 MHz versus 2400 MHz for the B370. The B390 therefore combines a modest clock uplift with a 20% resource increase, producing the 25% throughput advantage observed in the fill-rate and compute figures.

Power delivery differs substantially. The B370 has a 25 W TDP, while the B390 has an 80 W TDP. Both are integrated graphics processors (IGP) with no power connectors, a slot width of IGP, and a bus interface of IGP. The memory subsystem is entirely system-shared for both, with no dedicated VRAM, no fixed bus width, and bandwidth listed as system dependent. Display outputs are portable device dependent for both, meaning the database does not assign fixed monitor connectivity to either part.

The API support is identical: DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. Both are marked as active production parts, released on the same date, with the same predecessor (HD Graphics-WM) and no successor. No launch MSRP is recorded for either.

The Verdict

Based strictly on the specification data, the Intel Arc Pro B390 is the stronger part. It delivers 25% higher FP32 compute, 25% higher FP16 compute, 25% higher pixel fill rate, and 25% higher texture fill rate than the B370. It also boosts 100 MHz higher. The B390 achieves this with 20% more shading units, 20% more TMUs, 20% more ROPs, and 20% more RT cores. Every performance-relevant metric favors the B390.

The B370 counters with a 25 W TDP versus 80 W for the B390, a 3.2x reduction in power draw. For integrated graphics in portable devices, this power difference is significant. The B370 consumes far less energy while still providing 80% of the execution resources and 96% of the boost clock. The B390 is not a different architecture, it is a wider and faster implementation of the same Xe3-LPG design with a much higher power envelope.

The data indicates that system designers seeking maximum graphics throughput from a Panther Lake platform should choose the B390. Those prioritizing power efficiency in a thermally constrained portable chassis should choose the B370. Neither part has measured benchmark scores, so real-world performance parity cannot be confirmed, but the theoretical throughput gap is consistent and unambiguous.

FAQ

Q: Which GPU has higher FP32 compute performance?

A: The Intel Arc Pro B390 has 7.680 TFLOPS FP32 compute, which is 25% higher than the 6.144 TFLOPS of the Intel Arc Pro B370.

Q: How do the boost clocks compare?

A: The Arc Pro B390 boosts to 2500 MHz, while the Arc Pro B370 boosts to 2400 MHz, a 100 MHz difference in favor of the B390.

Q: Are the shading unit counts different?

A: Yes, the B390 has 1536 shading units, while the B370 has 1280 shading units, a 20% difference.

Q: Do both GPUs use the same architecture?

A: Yes, both use the Xe3-LPG architecture on the Panther Lake chip, manufactured on a 3 nm process at Intel.

Q: What is the TDP difference?

A: The Arc Pro B370 has a 25 W TDP, while the Arc Pro B390 has an 80 W TDP.

Q: Is there any dedicated video memory on either GPU?

A: No, both GPUs use system shared memory, with no dedicated VRAM, no fixed bus width, and bandwidth dependent on the system.

Where Each One Wins

The Arc Pro B390 wins on every raw throughput metric recorded in the database. Its 7.680 TFLOPS FP32 and 15.36 TFLOPS FP16 give it the compute edge for general-purpose GPU workloads, shader-heavy rendering, and any task that scales with shading unit count. The 120.0 GTexel/s texture rate and 60.00 GPixel/s pixel rate make it the stronger choice for texture-bound and rasterization-bound scenes. The 12 RT cores versus 10 give it a proportional advantage in ray tracing workloads, assuming the 20% resource scaling translates to real-world ray tracing performance.

The Arc Pro B370 wins on power efficiency. At 25 W, it uses 68.75% less power than the B390's 80 W while delivering 80% of the execution resources and 96% of the boost clock. For thermally constrained portable devices, this efficiency gap is the deciding factor. The B370 also shares the same API support, the same process node, and the same memory architecture, so it loses no feature compatibility.

The B390 is the performance pick for any scenario where power draw is not the limiting factor. The B370 is the efficiency pick for thin-and-light systems, fanless designs, or battery-sensitive workloads where the 25% throughput reduction is an acceptable trade for the 3.2x lower TDP.

Specification Differences

| Field | Intel Arc Pro B370 | Intel Arc Pro B390 |

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

| Boost Clock | 2400 MHz | 2500 MHz |

| Shading Units | 1280 | 1536 |

| TMUs | 40 | 48 |

| ROPs | 20 | 24 |

| RT Cores | 10 | 12 |

| Pixel Rate | 48.00 GPixel/s | 60.00 GPixel/s |

| Texture Rate | 96.00 GTexel/s | 120.0 GTexel/s |

| FP32 Compute | 6.144 TFLOPS | 7.680 TFLOPS |

| FP16 Compute | 12.29 TFLOPS (2:1) | 15.36 TFLOPS (2:1) |

| TDP | 25 W | 80 W |

All other fields match: both use the Panther Lake chip, Xe3-LPG architecture, 3 nm process, Intel foundry, 300 MHz base clock, system shared memory, IGP slot width, no power connectors, portable device dependent display outputs, DirectX 12 Ultimate (12_2), OpenGL 4.6, Vulkan 1.4, active production status, and the same release date. Neither part has a recorded launch MSRP, benchmark scores, or nearest rivals in the database.