Intel Arc Pro A60M vs Intel Arc Pro B65 Comparison

Intel Arc Pro A60M and Intel Arc Pro B65 represent two distinct generations of Intel’s professional mobile and desktop GPU lineup. The A60M is an Alchemist-generation part built on TSMC’s 6 nm process, while the B65 belongs to the Battlemage family on TSMC’s 5 nm node. The recorded data shows a substantial gap in raw specifications, memory capacity, and power requirements, making them suitable for different classes of workloads despite sharing the same DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4 API support. The A60M is an integrated graphics processor (IGP) with a 95 W TDP, while the B65 is a dual-slot discrete card with a 200 W TDP and a single 8-pin power connector. Neither part has a recorded launch MSRP in the database.

Where Each One Wins

The Intel Arc Pro A60M wins in scenarios where power envelope and physical footprint are the primary constraints. As an IGP with a 95 W TDP, it fits into portable or compact systems where a dual-slot card is not an option. Its 8 GB GDDR6 memory on a 128 bit bus delivers 256.0 GB/s of bandwidth, which is sufficient for moderate professional workloads that do not require large frame buffers. The A60M uses a PCIe 4.0 x16 interface, which is compatible with a wide range of existing platforms. Its display outputs are marked as portable device dependent, meaning the manufacturer of the host system determines the actual ports. This makes the A60M a flexible choice for mobile workstations or embedded designs where the GPU must be integrated into the motherboard rather than installed as an expansion card.

The Intel Arc Pro B65 wins decisively in raw compute and memory capacity. It uses the BMG-G21 chip with 19,600 million transistors on a 272 mm² die, compared to the A60M’s DG2-256 chip with 11,500 million transistors on a 269 mm² die. The B65 delivers 12.29 TFLOPS of FP32 performance versus 5.325 TFLOPS for the A60M, more than doubling the compute throughput. Its 32 GB GDDR6 memory on a 256 bit bus provides 608.0 GB/s of bandwidth, which is 2.375 times the A60M’s bandwidth. The B65 also has a significantly higher boost clock of 2400 MHz, matching its base clock, whereas the A60M boosts to 1300 MHz from a 900 MHz base. For workloads that depend on memory capacity, such as large model rendering, simulation, or multi-layer compositing, the B65 is the clear leader.

The B65 also wins on pixel and texture throughput. Its pixel rate of 192.0 GPixel/s is 2.3 times the A60M’s 83.20 GPixel/s, and its texture rate of 384.0 GTexel/s is 2.3 times the A60M’s 166.4 GTexel/s. These figures indicate that the B65 can handle higher resolution output and more complex texture filtering without becoming the bottleneck. The B65’s 20 ray tracing cores, compared to 16 on the A60M, give it an advantage in ray-traced professional visualization workloads, though the database does not record a specific benchmark delta for this comparison. The B65 also supports PCIe 5.0 x16, doubling the bus bandwidth available to the A60M’s PCIe 4.0 x16 connection, which matters for data transfer to and from system memory in large dataset workflows.

The Verdict

The data indicates that the Intel Arc Pro B65 is the stronger performer across every measured specification category. It has more shading units, texture mapping units, and ROPs: 2560 versus 2048, 160 versus 128, and 80 versus 64, respectively. Its FP32 throughput is 12.29 TFLOPS, which is 2.3 times the A60M’s 5.325 TFLOPS. Memory bandwidth is 608.0 GB/s versus 256.0 GB/s, a 2.375 times advantage. The B65’s 32 GB frame buffer is four times the capacity of the A60M’s 8 GB. For any workload where compute, memory, or display output capability matters, the B65 is the appropriate choice.

The Intel Arc Pro A60M remains relevant for a different class of system. Its 95 W TDP and IGP form factor mean it can be deployed in notebooks, compact desktops, or industrial systems where space and cooling are limited. The A60M’s 8 GB GDDR6 memory is adequate for entry-level professional tasks, and its 16 ray tracing cores provide basic RT acceleration. The A60M’s 50th percentile ranking among all GPUs in the database matches the B65’s percentile, but this ranking does not account for the fact that the A60M is a mobile-class part while the B65 is a desktop-class card. The choice between the two is not about performance parity; it is about system integration constraints. The B65 requires a dual-slot chassis and a 550 W suggested PSU, while the A60M does not have a recorded suggested PSU or power connector requirement.

Head-to-Head Benchmarks

The database does not list direct head-to-head benchmark scores for these two GPUs, so the comparison relies on the recorded specification deltas. The most significant gap is in FP32 compute. The B65’s 12.29 TFLOPS is 2.308 times the A60M’s 5.325 TFLOPS. In practical terms, a workload that takes one hour on the A60M would take approximately 26 minutes on the B65, assuming perfect scaling and no other bottlenecks. The FP16 figures follow the same ratio: the B65 delivers 24.58 TFLOPS versus the A60M’s 10.65 TFLOPS, again a 2.308 times difference. Both parts use a 2:1 FP16 to FP32 ratio, so the relative advantage is consistent across precision levels.

Memory bandwidth is the second largest differentiator. The B65’s 608.0 GB/s is 2.375 times the A60M’s 256.0 GB/s. This matters for memory-bound workloads such as high-resolution texture streaming, large geometry buffers, or multi-pass rendering where data must be read and written repeatedly. The B65’s 32 GB capacity also allows it to hold datasets that simply cannot fit within the A60M’s 8 GB frame buffer. For example, a 3D scene with high-resolution textures exceeding 8 GB would cause the A60M to spill to system memory, while the B65 would keep it resident on the GPU.

Pixel throughput shows a 2.308 times advantage for the B65: 192.0 GPixel/s versus 83.20 GPixel/s. Texture throughput is also 2.308 times higher: 384.0 GTexel/s versus 166.4 GTexel/s. These ratios match the FP32 ratio because the B65 has exactly 1.25 times the shading units, TMUs, and ROPs of the A60M, but it runs at a much higher clock. The B65’s 2400 MHz boost clock is 1.846 times the A60M’s 1300 MHz boost clock. When combined with the 25% increase in execution units, the result is the observed 2.308 times performance ratio. The B65 also has 20 ray tracing cores versus 16, a 25% increase, but the clock advantage applies equally to those cores.

Transistor density is another differentiating factor. The B65 packs 19,600 million transistors into a 272 mm² die, yielding 72.1 million transistors per square millimeter. The A60M has 11,500 million transistors on a 269 mm² die, yielding 42.8 million transistors per square millimeter. The B65’s 5 nm process allows for 68.5% higher transistor density than the A60M’s 6 nm process. Despite the higher density, the B65’s die is only 3 mm² larger, which means the B65 uses its silicon area much more efficiently.

FAQ

Q: Which GPU has more memory?

A: The Intel Arc Pro B65 has 32 GB of GDDR6 memory, which is four times the 8 GB found on the Intel Arc Pro A60M.

Q: What is the FP32 performance difference?

A: The B65 delivers 12.29 TFLOPS of FP32 compute, while the A60M delivers 5.325 TFLOPS. The B65 is 2.308 times faster in this metric.

Q: Do they use the same architecture?

A: No. The A60M uses the Xe-HPG architecture with the DG2-256 chip from the Alchemist generation, while the B65 uses the Xe2-HPG architecture with the BMG-G21 chip from the Battlemage generation.

Q: What are the power requirements?

A: The A60M has a 95 W TDP and is an IGP with no power connector listed. The B65 has a 200 W TDP, uses a single 8-pin power connector, and has a suggested PSU of 550 W.

Q: Which GPU supports a newer PCIe interface?

A: The B65 supports PCIe 5.0 x16, while the A60M supports PCIe 4.0 x16.

Q: What is the memory bandwidth for each?

A: The A60M has 256.0 GB/s of bandwidth on a 128 bit bus, and the B65 has 608.0 GB/s on a 256 bit bus. The B65 provides 2.375 times the bandwidth.

Architecture Differences

The two GPUs come from different architectural generations. The A60M is based on Xe-HPG, Intel’s first-generation high-performance graphics architecture, using the DG2-256 chip. The B65 uses Xe2-HPG, the second-generation architecture, with the BMG-G21 chip. The manufacturing process differs as well: the A60M is built on TSMC’s 6 nm node, while the B65 uses TSMC’s 5 nm node. This process change contributes to the B65’s higher transistor density of 72.1 million transistors per square millimeter, compared to the A60M’s 42.8 million. The B65 also has a higher transistor count of 19,600 million versus 11,500 million, despite a nearly identical die size of 272 mm² versus 269 mm².

The execution unit configuration scales with the architecture generation. The A60M has 2048 shading units, 128 TMUs, 64 ROPs, and 16 ray tracing cores. The B65 has 2560 shading units, 160 TMUs, 80 ROPs, and 20 ray tracing cores. This represents a 25% increase in each execution resource category. However, the clock speed difference is larger: the B65 runs at a constant 2400 MHz for both base and boost, while the A60M has a 900 MHz base and a 1300 MHz boost. The B65’s base clock is 2.667 times the A60M’s base clock and 1.846 times its boost clock.

Memory architecture also differs fundamentally. The A60M uses an 8 GB GDDR6 configuration with a 128 bit bus, resulting in 256.0 GB/s of bandwidth. The B65 uses 32 GB GDDR6 with a 256 bit bus, resulting in 608.0 GB/s. The memory clock is higher on the B65 as well: 2375 MHz with 19 Gbps effective data rate, versus 2000 MHz with 16 Gbps effective on the A60M. The B65’s larger bus width and higher memory clock combine to produce its 2.375 times bandwidth advantage.

Specification Differences

The recorded specifications differ across nearly every field. The A60M uses the DG2-256 chip on a 6 nm TSMC process with 11,500 million transistors and a 269 mm² die. The B65 uses the BMG-G21 chip on a 5 nm TSMC process with 19,600 million transistors and a 272 mm² die. Clock speeds: the A60M has a 900 MHz base and 1300 MHz boost, while the B65 has a 2400 MHz base and 2400 MHz boost. Memory: the A60M has 8 GB GDDR6 on a 128 bit bus with 256.0 GB/s bandwidth and a 2000 MHz memory clock (16 Gbps effective). The B65 has 32 GB GDDR6 on a 256 bit bus with 608.0 GB/s bandwidth and a 2375 MHz memory clock (19 Gbps effective).

Compute resources: the A60M has 2048 shading units, 128 TMUs, 64 ROPs, and 16 ray tracing cores. The B65 has 2560 shading units, 160 TMUs, 80 ROPs, and 20 ray tracing cores. Pixel rate: 83.20 GPixel/s for the A60M versus 192.0 GPixel/s for the B65. Texture rate: 166.4 GTexel/s versus 384.0 GTexel/s. FP32: 5.325 TFLOPS versus 12.29 TFLOPS. FP16: 10.65 TFLOPS versus 24.58 TFLOPS, both with a 2:1 ratio. Power: the A60M has a 95 W TDP, while the B65 has a 200 W TDP. The B65 requires a single 8-pin power connector and a 550 W suggested PSU, while the A60M has no power connector or PSU listed. Form factor: the A60M is an IGP, and the B65 is dual-slot. Bus interface: the A60M uses PCIe 4.0 x16, and the B65 uses PCIe 5.0 x16. Display outputs: the A60M is portable device dependent, and the B65 has 4x DisplayPort 2.1. Both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The A60M was released on 2023-06-05, and the B65 is dated 2026-03-31.