Intel Arc A350M vs Intel Arc A370M Comparison

Head-to-Head Benchmarks

The recorded data shows a clear and consistent performance advantage for the Intel Arc A370M across both benchmark workloads. In the Geekbench OpenCL test, the A370M scores 29,676 against the A350M's 24,546, a decisive 20.9% lead. The Vulkan result narrows the gap slightly but still favors the A370M: 28,673 versus 24,747, a 15.9% advantage. These are not marginal differences; the A370M wins both recorded head-to-head contests outright.

Context from the database's nearest-rival comparisons reinforces the significance of this spread. The A370M's average benchmark score sits at 29,175, placing it within 0.1% of the AMD Radeon RX Vega M GH (29,197) and the AMD FirePro W8000 (29,211). It also edges out the AMD Radeon RX 470 (28,996) by 0.6% and the AMD Radeon RX 6800M (28,874) by 1%. By contrast, the A350M's average score of 24,647 places it just 0.4% behind the AMD Radeon RX 590 (24,744) and 0.5% behind the NVIDIA RTX A5000 Mobile (24,763). The A350M does lead the AMD Radeon RX 6600 XT (24,442) by 0.8% and the NVIDIA GeForce GTX 1630 (24,277) by 1.5%, but its overall bracket sits roughly 4,500 points lower than the A370M's.

The percentile rankings tell the same story. The A370M lands in the 74th percentile of all GPUs in the database, while the A350M reaches only the 70th percentile. That 4-point gap may sound small, but it represents a meaningful tier separation when combined with the raw score deltas. In practical terms, the A370M delivers roughly one-fifth more OpenCL throughput and one-sixth more Vulkan throughput than the A350M, which is a substantial generational step within the same product family.

Architecture Differences

Both GPUs share the same fundamental design: the DG2-128 chip built on Intel's Xe-HPG architecture, fabricated on TSMC's 6 nm process. Each chip contains 7,200 million transistors on a 157 mm² die, yielding a transistor density of 45.9 million per square millimeter. Both are classified under the Alchemist (Arc 3 Mobile) generation and carry identical API support: DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

The divergence begins with execution resources. The A370M enables 1,024 shading units, 64 texture mapping units (TMUs), and 32 raster operation units (ROPs). The A350M ships with 768 shading units, 48 TMUs, and 24 ROPs. That means the A370M has 33.3% more shaders, 33.3% more TMUs, and 33.3% more ROPs than the smaller part. Ray tracing cores follow the same pattern: 8 on the A370M versus 6 on the A350M, a 33.3% increase.

Clock behavior adds another layer of separation, but not in a simple linear way. The A370M has a higher base clock at 1550 MHz, while the A350M starts at 1150 MHz. However, the boost clocks invert this relationship: the A350M boosts to 2200 MHz, exceeding the A370M's 2050 MHz boost. This suggests the A350M is tuned for bursty workloads where its lower power envelope allows a high temporary ceiling, while the A370M relies on a higher sustained floor plus additional execution units. Memory clocks are identical: 1750 MHz with 14 Gbps effective data rate in both cases.

The raw throughput figures reflect the combination of clocks and core counts. The A370M achieves 65.60 GPixel/s pixel fill rate and 131.2 GTexel/s texture fill rate. The A350M produces 52.80 GPixel/s and 105.6 GTexel/s. For FP32 compute, the A370M reaches 4.198 TFLOPS, while the A350M manages 3.379 TFLOPS. FP16 numbers scale proportionally at a 2:1 ratio: 8.397 TFLOPS for the A370M and 6.758 TFLOPS for the A350M.

Power consumption differs as well. The A370M carries a 35 W TDP, while the A350M is rated at 25 W. Both are integrated into the IGP slot width with no dedicated power connectors, and both rely on portable-device-dependent display outputs rather than fixed panel configurations.

Where Each One Wins

The A370M wins both recorded benchmark tests, so the database offers no workload where the A350M comes out ahead in raw performance. That does not mean the A350M lacks purpose; it simply means its advantages are not measured in FPS or compute scores.

The A350M's primary strength is efficiency. At 25 W TDP, it draws 10 W less than the A370M's 35 W rating. For thin-and-light laptops with constrained thermal budgets, that power delta can be the deciding factor between a usable chassis and one that throttles under load. The A350M also posts a higher boost clock (2200 MHz versus 2050 MHz), which may help in short, bursty tasks that do not saturate all cores. If a workload is latency-sensitive and lightly threaded, the A350M's higher boost ceiling could keep it competitive despite fewer execution units.

The A370M, by contrast, wins everywhere the database measures performance. Its 20.9% OpenCL advantage and 15.9% Vulkan advantage translate directly to better general-purpose compute and graphics API performance. The 33.3% increase in shading units, TMUs, ROPs, and ray tracing cores makes it the clear choice for rendering, ray-traced effects, and any parallel compute workload. Its higher base clock (1550 MHz versus 1150 MHz) also means sustained full-load performance is likely more stable, since the A370M does not need to rely on boost headroom to reach competitive speeds.

Users who prioritize raw capability in a compact mobile form factor should select the A370M. Users who prioritize minimum power draw above all else, and who can accept lower absolute performance, would look to the A350M. The data does not show any scenario where the A350M outperforms the A370M in measured throughput, so the choice reduces to whether the 10 W TDP savings justify the performance penalty.

Specification Differences

The following fields differ between the Intel Arc A370M and Intel Arc A350M. All other recorded specifications are identical.

| Field | Intel Arc A370M | Intel Arc A350M |

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

| Base clock | 1550 MHz | 1150 MHz |

| Boost clock | 2050 MHz | 2200 MHz |

| Shading units | 1024 | 768 |

| TMUs | 64 | 48 |

| ROPs | 32 | 24 |

| Ray tracing cores | 8 | 6 |

| Pixel rate | 65.60 GPixel/s | 52.80 GPixel/s |

| Texture rate | 131.2 GTexel/s | 105.6 GTexel/s |

| FP32 | 4.198 TFLOPS | 3.379 TFLOPS |

| FP16 | 8.397 TFLOPS (2:1) | 6.758 TFLOPS (2:1) |

| TDP | 35 W | 25 W |

| Geekbench OpenCL | 29676 | 24546 |

| Geekbench Vulkan | 28673 | 24747 |

| Average benchmark score | 29175 | 24647 |

| Percentile vs all GPUs | 74 | 70 |

Identical items include the DG2-128 chip, Xe-HPG architecture, 6 nm process, TSMC foundry, 7,200 million transistors, 157 mm² die size, 45.9M/mm² transistor density, 4 GB GDDR6 memory, 64-bit bus width, 112.0 GB/s bandwidth, 1750 MHz memory clock with 14 Gbps effective rate, PCIe 4.0 x8 interface, IGP slot width, DirectX 12 Ultimate (12_2), OpenGL 4.6, Vulkan 1.4, portable-device-dependent display outputs, end-of-life production status, and a release date of 2022-03-29.

FAQ

Q: Which GPU is faster in the recorded benchmarks?

A: The Intel Arc A370M wins both recorded tests. It scores 29,676 in Geekbench OpenCL versus 24,546 for the A350M, and 28,673 in Geekbench Vulkan versus 24,747.

Q: How much larger is the A370M's performance advantage in OpenCL compared to Vulkan?

A: The A370M leads by 20.9% in Geekbench OpenCL and by 15.9% in Geekbench Vulkan.

Q: Do the two GPUs use the same physical chip?

A: Yes, both use the DG2-128 chip on the Xe-HPG architecture, built on TSMC's 6 nm process with 7,200 million transistors on a 157 mm² die.

Q: What is the difference in shading units?

A: The A370M has 1,024 shading units, while the A350M has 768, a 33.3% increase for the A370M.

Q: Does the A350M have a higher boost clock than the A370M?

A: Yes. The A350M boosts to 2200 MHz boost clock, which is higher than the A370M's 2050 MHz boost clock. However, the A370M has the higher base clock at 1550 MHz.

Q: How do their power ratings compare?

A: The A370M is rated at 35 W TDP, while the A350M is rated at 25 W TDP, a 10 W difference favoring the A350M.

Q: Where does the A370M rank among all GPUs in the database by percentile, and where does the A350M rank?

A: The A370M ranks in the 74th percentile, while the A350M ranks in the 70th percentile.

Q: What average benchmark scores does each GPU report?

A: The A370M averages 29,175, while the A350M averages 24,647.

The Verdict

The Intel Arc A370M is the superior product on every measured performance axis in this database. It wins both head-to-head benchmarks, posts a higher average score (29,175 versus 24,647), and ranks higher in the overall percentile distribution (74 versus 70). The A370M's 1024 shading units, 64 TMUs, 32 ROPs, and 8 ray tracing cores give it a 33.3% resource advantage over the A350M across each of those categories. Its lower boost clock (2050 MHz versus 2200 MHz) does not prevent it from outpacing the A350M in either OpenCL or Vulkan workloads, which indicates that the extra execution units more than compensate for the clock deficit.

The A350M is not without a rationale. Its 25 W TDP makes it a better fit for ultraportable designs where every watt matters, and its higher boost clock could help in short burst workloads. But the database does not show any benchmark where the A350M wins outright. For any user who wants the strongest available mobile Arc 3 performance, the A370M is the only choice supported by the data. Users who absolutely require the lower power envelope will accept a 20.9% OpenCL and 15.9% Vulkan penalty as the cost of that efficiency.