Intel Arc B370 vs Intel Arc Graphics 48EU Mobile Comparison

The Verdict

The Intel Arc B370 and Intel Arc Graphics 48EU Mobile are both integrated graphics solutions from Intel, but they serve distinctly different tiers of mobile performance. The recorded data shows the Arc B370 is the decisive performer, with a 3DMark Steel Nomad DX12 score of 1184, while the Arc Graphics 48EU Mobile has no recorded benchmark score in the database, placing it in the 50th percentile among all GPUs compared to the B370's 5th percentile. That percentile gap requires careful interpretation: the B370's lower percentile reflects the fact that the 3DMark Steel Nomad test is a demanding modern workload, and the database includes many desktop discrete GPUs that score far higher, whereas the 48EU Mobile's 50th percentile is a default placement for a part with no direct measurements.

For users selecting between these two, the data points clearly to the Arc B370 for any task involving modern 3D rendering, particularly DirectX 12 Ultimate workloads. The B370's 1280 shading units, 40 texture mapping units, 20 raster operation units, and 10 ray tracing cores provide a hardware foundation that the 48EU Mobile lacks entirely on the ray tracing front. The 48EU Mobile, with 384 shading units, 24 TMUs, and 8 ROPs, is positioned for lighter graphics duties. Its predecessor is listed as HD Graphics-M, indicating it represents a baseline integrated graphics evolution rather than a performance-oriented part.

The B370 is the correct choice for a system expected to run contemporary games or GPU-accelerated creative applications, even at modest settings. The 48EU Mobile suits systems where graphics are secondary, such as office productivity, media playback, or light 2D workloads. The B370's higher boost clock of 2400 MHz against the 48EU's 1800 MHz, combined with its larger execution resource pool, creates a performance hierarchy that the benchmark data supports, even though the 48EU has no direct score to quantify the gap.

Where Each One Wins

The Arc B370 wins in every measurable category where hardware specifications predict performance. Its pixel rate of 48.00 GPixel/s is more than three times the 48EU Mobile's 14.40 GPixel/s, meaning the B370 can fill the framebuffer substantially faster, which directly benefits high-resolution rendering and heavy post-processing effects. The texture rate of 96.00 GTexel/s versus 43.20 GTexel/s gives the B370 a 2.2x advantage in texture fetch throughput, which matters for games with detailed surfaces and for compute workloads that sample textures repeatedly.

The FP32 compute throughput of 6.144 TFLOPS on the B370 dwarfs the 48EU Mobile's 1,382.4 GFLOPS, a 4.4x gap. This directly translates to faster shader execution, physics simulation, and any general-purpose GPU compute task. The FP16 figure follows the same pattern: 12.29 TFLOPS on the B370 versus 2.765 TFLOPS on the 48EU Mobile, again a roughly 4.4x difference. For applications that use FP16 for machine learning inference or image processing, the B370 provides a much larger compute envelope.

The B370 also wins on API support. It lists DirectX 12 Ultimate (12_2), which includes features like ray tracing and mesh shaders, while the 48EU Mobile peaks at DirectX 12 (12_1), which lacks the full 12_2 feature set. Both parts support OpenGL 4.6 and Vulkan 1.4, so cross-platform APIs are equally covered, but the B370's DX12 Ultimate capability is a clear functional advantage for modern Windows titles.

The 48EU Mobile does have one specification where it leads: TDP. The B370 is rated at 25 W, while the 48EU Mobile is rated at 28 W. This is a counterintuitive finding: the smaller, more capable B370 consumes less power according to the database, which suggests the newer 3 nm process node delivers efficiency gains. For a thin-and-light laptop chassis, the B370's lower TDP combined with higher performance makes it the superior engineering choice, assuming thermal and power delivery designs accommodate it.

Architecture Differences

The two GPUs come from different architecture generations and process nodes. The Intel Arc B370 uses the Xe3-LPG architecture on a 3 nm process, built on the Panther Lake chip. The Intel Arc Graphics 48EU Mobile uses the Xe-LPG architecture on a 10 nm process, built on the Meteor Lake chip. The B370 belongs to the Arc Graphics-M (Panther Lake) generation, while the 48EU Mobile belongs to the Arc Graphics-M (Meteor Lake) generation. This generational split explains the substantial specification gaps.

The B370's execution resources are scaled up considerably: 1280 shading units versus 384 on the 48EU Mobile, 40 TMUs versus 24, 20 ROPs versus 8, and 10 ray tracing cores versus none listed for the 48EU Mobile. The B370 also has a higher boost clock at 2400 MHz versus 1800 MHz, while both share the same 300 MHz base clock. The B370's clock advantage compounds its execution resource advantage, so the real-world performance gap is likely larger than the raw unit counts suggest.

Memory configuration is identical in structure: both use System Shared memory, meaning they rely on the host system's RAM rather than dedicated VRAM. The bus width is System Shared for both, and bandwidth is System Dependent for both. This means the memory subsystem performance depends entirely on the system memory configuration, and neither GPU has a dedicated memory advantage in the specifications. The practical implication is that dual-channel high-speed system memory will benefit both parts equally, and the B370's larger compute resources may be partially constrained by shared memory bandwidth in memory-intensive workloads.

The B370's bus interface is listed as IGP, meaning integrated graphics processor, while the 48EU Mobile's bus interface is listed as Ring Bus. This indicates a difference in how the GPUs connect to the rest of the chip. The Ring Bus interface on the 48EU Mobile suggests a more tightly integrated on-die connection within the Meteor Lake package, while the B370's IGP designation on Panther Lake likely reflects a similar integrated placement but with a different internal topology. Neither interface is quantified in terms of bandwidth, so the data does not permit a direct comparison of interconnect performance.

Both GPUs use the same display outputs, listed as Portable Device Dependent, meaning the actual ports depend on the laptop or mobile device design. The B370's power connectors are listed as None, while the 48EU Mobile has no power connector data listed, which is consistent with both being integrated parts that draw power from the motherboard rather than external connectors. The B370's slot width is IGP, and the 48EU Mobile's slot width is also IGP, confirming both are solder-down integrated solutions.

The B370 was released on 2026-01-26, while the 48EU Mobile was released on 2023-12-13. The 48EU Mobile has a predecessor listed as HD Graphics-M, while the B370 has no predecessor listed. The 48EU Mobile has no successor listed, and neither part has a successor. Production status for both is Active. The B370's process node advantage, 3 nm versus 10 nm, is a significant architectural difference that explains how the B370 achieves higher clock speeds and more execution units while maintaining a lower TDP of 25 W versus 28 W.

FAQ

Q: Which GPU has better raw compute performance?

A: The Arc B370 delivers 6.144 TFLOPS FP32, which is 4.4 times the 1,382.4 GFLOPS of the Arc Graphics 48EU Mobile. The FP16 figures show the same ratio: 12.29 TFLOPS versus 2.765 TFLOPS.

Q: Does the Arc B370 support ray tracing?

A: Yes, the B370 has 10 ray tracing cores and supports DirectX 12 Ultimate (12_2). The Arc Graphics 48EU Mobile has no ray tracing cores listed and supports only DirectX 12 (12_1).

Q: What is the clock speed difference?

A: Both GPUs have a 300 MHz base clock. The Arc B370 boosts to 2400 MHz, while the Arc Graphics 48EU Mobile boosts to 1800 MHz. The B370's boost clock is 600 MHz higher.

Q: Which GPU consumes less power?

A: The Arc B370 is rated at 25 W TDP, while the Arc Graphics 48EU Mobile is rated at 28 W TDP. The B370 uses less power despite having significantly more execution resources.

Q: What memory do these GPUs use?

A: Both use System Shared memory, meaning they share the host system's RAM. The memory type is System Shared for both, the bus width is System Shared for both, and bandwidth is System Dependent for both.

Q: How do the shading unit counts compare?

A: The Arc B370 has 1280 shading units, while the Arc Graphics 48EU Mobile has 384 shading units. The B370 also has 40 TMUs versus 24, and 20 ROPs versus 8.

Head-to-Head Benchmarks

The database contains no direct head-to-head benchmark entries between these two GPUs, and the wins counter shows zero wins for either part in a direct comparison. However, the Arc B370 has a single recorded benchmark result: a 3DMark Steel Nomad DX12 score of 1184. The Arc Graphics 48EU Mobile has no recorded benchmark scores at all, with an average benchmark score of 0. This absence of data for the 48EU Mobile means the comparison must rely on the B370's absolute score and its position relative to other GPUs in the database.

The B370's 3DMark Steel Nomad DX12 score of 1184 places it near several rival GPUs. The nearest rival is the ATI Mobility Radeon HD 5570 with an average score of 1186, a delta of -0.2 percent, meaning the B370 is essentially tied with that older mobile part. The ATI Radeon HD 5770 scores 1190, putting the B370 0.5 percent behind. The AMD Radeon HD 7650M scores 1192, putting the B370 0.7 percent behind. The AMD FirePro M2000 scores 1168, and the B370 is 1.4 percent ahead of that part. These deltas are all within a narrow band, indicating the B370's measured performance clusters with a group of GPUs from a much older era of mobile and desktop graphics.

The B370's 5th percentile ranking across all GPUs in the database reflects this clustering: the 3DMark Steel Nomad test is a demanding modern benchmark, and the B370's score of 1184 places it below the vast majority of recorded GPUs. This percentile is not a commentary on the B370's quality but rather on the benchmark's difficulty and the database's composition, which includes many high-end discrete cards. The 48EU Mobile's 50th percentile with no benchmark score is a default mid-pack placement, not a measured result, so it should not be interpreted as the 48EU Mobile outperforming the B370 in any test.

The specification comparison provides the clearest head-to-head picture. The B370's pixel rate of 48.00 GPixel/s is 3.3 times the 48EU Mobile's 14.40 GPixel/s. The texture rate of 96.00 GTexel/s on the B370 is 2.2 times the 48EU Mobile's 43.20 GTexel/s. The FP32 compute of 6.144 TFLOPS versus 1,382.4 GFLOPS is a 4.4x gap. These three throughput metrics are the most direct predictors of rasterization performance, and all three favor the B370 by a wide margin.

The shading unit count difference of 1280 versus 384 produces a 3.3x ratio, and the TMU count difference of 40 versus 24 produces a 1.7x ratio, while the ROP count difference of 20 versus 8 produces a 2.5x ratio. The B370's higher boost clock of 2400 MHz versus 1800 MHz adds an additional multiplier on top of these resource ratios. The combination of more units and higher clocks means the B370's theoretical peak throughput is substantially higher across every pipeline stage.

The B370's ray tracing cores, 10 of them, represent a capability the 48EU Mobile does not offer at all. DirectX 12 Ultimate support on the B370 enables hardware-accelerated ray tracing in supported titles, while the 48EU Mobile's DirectX 12 (12_1) support cannot access those features. This is not a performance difference but a feature difference, and it matters for any workload that requires ray-traced effects.

The process node difference, 3 nm for the B370 versus 10 nm for the 48EU Mobile, explains the B370's ability to pack more execution units and run at a higher boost clock while still maintaining a lower TDP of 25 W versus 28 W. The B370's release date of 2026-01-26 is later than the 48EU Mobile's 2023-12-13, reflecting the newer process and architecture generation.

The B370's FP16 throughput of 12.29 TFLOPS (2:1) versus the 48EU Mobile's 2.765 TFLOPS (2:1) shows the same 4.4x ratio as FP32, indicating the 2:1 FP16 ratio is consistent across both parts. This matters for workloads that can use reduced precision, such as certain machine learning inference tasks or image processing algorithms.

For the 48EU Mobile, the absence of ray tracing cores and the lower DirectX version are the most consequential architectural limitations. The 48EU Mobile's 384 shading units and 8 ROPs are appropriate for modest 1080p gaming at low settings or for basic video playback, but the data does not support any scenario where the 48EU Mobile outperforms the B370. The B370's measured 3DMark Steel Nomad score, even with its low percentile ranking, confirms it can complete a demanding modern DX12 workload, while the 48EU Mobile has no recorded result to indicate it can do the same. All recorded data points to the Arc B370 as the stronger part in every performance category.