AMD Instinct MI350P vs Intel Arc Pro B60 Comparison

The Verdict

The AMD Instinct MI350P and Intel Arc Pro B60 serve fundamentally different purposes, and the recorded data supports a clear separation. The MI350P is an enterprise compute accelerator with no display outputs, aimed at high-throughput workloads. The Arc Pro B60 is a professional graphics card with four mini-DisplayPort 2.1 outputs, designed for workstation visualization and rendering. Benchmark results are only recorded for the Arc Pro B60, which holds a 20th percentile position among all GPUs with an average benchmark score of 3182. The MI350P has no recorded benchmarks, a 50th percentile position, and an average score of 0, indicating the database has not yet captured performance data for this part. Therefore, the verdict rests on architectural positioning: choose the MI350P for compute-intensive tasks requiring massive memory capacity and bandwidth, and choose the Arc Pro B60 for professional graphics workloads needing display output and API support.

Where Each One Wins

The Arc Pro B60 wins in every measurable benchmark category because it is the only one with recorded scores. Its 3DMark Steel Nomad DX12 score is 2646, its Passmark G3D score is 14580, and its Passmark GPU Compute score is 7029. The MI350P has no benchmark entries, so no direct comparison can be made from measured data. However, the MI350P wins on raw specifications that matter for compute acceleration: 144 GB of HBM3e memory versus 24 GB of GDDR6, 8.19 TB/s memory bandwidth versus 456.0 GB/s, and 36.04 TFLOPS FP32 versus 12.29 TFLOPS FP32. The MI350P also offers 8192 shading units, 512 TMUs, and a 8192-bit memory bus, all of which dwarf the Arc Pro B60's 2560 shading units, 160 TMUs, and 192-bit bus. For workloads that fit within the Arc Pro B60's 24 GB frame buffer, it delivers professional graphics features such as DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4 support, none of which the MI350P provides. The data indicates that the MI350P targets data-center-scale compute, while the Arc Pro B60 targets desktop workstation graphics.

Architecture Differences

The MI350P uses the CDNA 4.0 architecture on a 3 nm TSMC process, while the Arc Pro B60 uses the Xe2-HPG architecture on a 5 nm TSMC process. The MI350P's chip is the MI350 128CU, with 73,000 million transistors on a 1190 mm² die, yielding a density of 61.3 million transistors per square millimeter. The Arc Pro B60's chip is the BMG-G21, with 19,600 million transistors on a 272 mm² die, yielding a higher density of 72.1 million transistors per square millimeter. This means the Intel chip packs more transistors per area despite being on a larger process node. The MI350P's memory is HBM3e across an 8192-bit bus, whereas the Arc Pro B60 uses GDDR6 across a 192-bit bus. The MI350P has no ROPs and a 0 MPixel/s pixel rate, confirming it is not designed for rasterization. The Arc Pro B60 has 80 ROPs and a 192.0 GPixel/s pixel rate. The MI350P also lacks ray tracing cores, while the Arc Pro B60 includes 20 RT cores. The MI350P has no API support for DirectX, OpenGL, or Vulkan, while the Arc Pro B60 supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The MI350P's FP16 performance matches its FP32 at 36.04 TFLOPS (1:1 ratio), while the Arc Pro B60 doubles its FP32 rate for FP16 at 24.58 TFLOPS (2:1 ratio). The MI350P uses a PCIe 5.0 x16 interface, while the Arc Pro B60 uses PCIe 5.0 x8. The MI350P requires a 600 W TDP and a 1000 W suggested PSU, while the Arc Pro B60 draws 200 W and suggests a 550 W PSU.

FAQ

Q: Which card has a higher FP32 compute throughput?

A: The AMD Instinct MI350P delivers 36.04 TFLOPS FP32, which is nearly three times the 12.29 TFLOPS of the Intel Arc Pro B60.

Q: Does the AMD Instinct MI350P support display outputs?

A: No, the MI350P has no display outputs. The Intel Arc Pro B60 provides four mini-DisplayPort 2.1 outputs.

Q: What is the memory capacity difference?

A: The MI350P has 144 GB of HBM3e memory, while the Arc Pro B60 has 24 GB of GDDR6 memory. That is a six-fold difference in capacity.

Q: How do the benchmark scores compare?

A: Only the Arc Pro B60 has recorded benchmark scores. Its Passmark G3D score is 14580, its Passmark GPU Compute score is 7029, and its 3DMark Steel Nomad DX12 score is 2646. The MI350P has no recorded benchmark scores.

Q: Which card has a smaller physical footprint?

A: The Arc Pro B60 measures 167 mm in length and 69 mm in height, while the MI350P measures 267 mm in length and 111 mm in height. Both are dual-slot cards with a 40 mm width.

Q: What process nodes are used?

A: The MI350P uses a 3 nm TSMC process, while the Arc Pro B60 uses a 5 nm TSMC process. The Arc Pro B60 has a higher transistor density at 72.1 million per square millimeter versus 61.3 million for the MI350P.

Head-to-Head Benchmarks

Since the head-to-head benchmark array is empty and the MI350P has no individual benchmark scores, the only quantitative comparison comes from the Arc Pro B60's recorded results relative to its nearest rivals. The Arc Pro B60's average benchmark score is 3182. Its closest competitor is the NVIDIA Quadro P1000 with an average score of 3163, a delta of 0.6 percent in favor of the Arc Pro B60. The next rival is the NVIDIA GeForce GT 640 with an average score of 3210, where the Arc Pro B60 trails by 0.9 percent. The NVIDIA GeForce 920M scores 3287, putting the Arc Pro B60 behind by 3.2 percent. The NVIDIA GeForce RTX 5080 SUPER scores 3075, and the Arc Pro B60 leads it by 3.5 percent. These deltas are small, indicating that the Arc Pro B60 sits in a narrow performance band among these four GPUs. In individual tests, the Arc Pro B60 shows its highest relative strength in Passmark G3D at 14580 and Passmark GPU Compute at 7029. Its lowest scores are Passmark DirectX 10 at 61 and Passmark DirectX 12 at 76. The Passmark DirectX 11 score is 122, DirectX 9 is 179, and G2D is 763. The 3DMark Steel Nomad DX12 score of 2646 rounds out the recorded data. The MI350P's absence from these benchmarks means its performance cannot be ranked against these rivals or the Arc Pro B60 from measured data alone.

Specification Differences

The two cards differ across nearly every specification category. Process node: 3 nm for the MI350P versus 5 nm for the Arc Pro B60. Transistor count: 73,000 million versus 19,600 million. Die size: 1190 mm² versus 272 mm². Transistor density: 61.3 million per mm² versus 72.1 million per mm². Base clock: 1000 MHz versus 2000 MHz. Boost clock: 2200 MHz versus 2400 MHz. Memory clock: 2000 MHz (8 Gbps effective) versus 2375 MHz (19 Gbps effective). Memory size: 144 GB versus 24 GB. Memory type: HBM3e versus GDDR6. Memory bus width: 8192 bit versus 192 bit. Memory bandwidth: 8.19 TB/s versus 456.0 GB/s. Shading units: 8192 versus 2560. TMUs: 512 versus 160. ROPs: 0 versus 80. RT cores: none versus 20. Pixel rate: 0 MPixel/s versus 192.0 GPixel/s. Texture rate: 1,126.4 GTexel/s versus 384.0 GTexel/s. FP32: 36.04 TFLOPS versus 12.29 TFLOPS. FP16: 36.04 TFLOPS (1:1) versus 24.58 TFLOPS (2:1). TDP: 600 W versus 200 W. Power connectors: 1x 16-pin versus 1x 8-pin. Suggested PSU: 1000 W versus 550 W. Bus interface: PCIe 5.0 x16 versus PCIe 5.0 x8. Display outputs: none versus 4x mini-DisplayPort 2.1. API support: none versus DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. Dimensions: 267 mm length, 111 mm height, 40 mm width versus 167 mm length, 69 mm height, 40 mm width. Production status: not listed versus Active. Release dates: the MI350P is dated 2026-05-06, while the Arc Pro B60 is dated 2025-09-04. The Arc Pro B60 has a launch MSRP of 499 USD, while the MI350P has no launch MSRP recorded. The MI350P's predecessor is Radeon Instinct, while the Arc Pro B60 has no predecessor listed. Neither has a recorded successor.