Intel Arc Pro B60 Dual vs NVIDIA RTX 5000 Ada Generation Comparison

FAQ

Q: How do the two cards compare in raw compute performance?

A: The NVIDIA RTX 5000 Ada Generation delivers 65.28 TFLOPS FP32 and 65.28 TFLOPS FP16 (1:1), while the Intel Arc Pro B60 Dual provides 12.29 TFLOPS FP32 and 24.58 TFLOPS FP16 (2:1). The NVIDIA card holds a 5.3x advantage in FP32 throughput.

Q: Which card offers more memory and bandwidth?

A: The RTX 5000 Ada has 32 GB of GDDR6 on a 256-bit bus with 576.0 GB/s bandwidth. The Arc Pro B60 Dual has 24 GB of GDDR6 on a 192-bit bus with 456.0 GB/s bandwidth. NVIDIA leads by 8 GB and 120 GB/s.

Q: What are the power requirements for each card?

A: The Arc Pro B60 Dual has a 400 W TDP with a suggested 800 W PSU. The RTX 5000 Ada has a 250 W TDP with a suggested 600 W PSU. NVIDIA consumes 150 W less and requires a smaller power supply.

Q: Which card has more RT and tensor cores?

A: The RTX 5000 Ada has 100 RT cores and 400 tensor cores. The Arc Pro B60 Dual has 20 RT cores and no tensor cores listed. The NVIDIA card offers 5x the RT cores.

Q: How do the display outputs differ?

A: The Arc Pro B60 Dual provides 4x mini-DisplayPort 2.1, while the RTX 5000 Ada provides 4x DisplayPort 1.4a. Intel supports the newer DisplayPort standard.

Q: What is the interface and slot configuration?

A: The Arc Pro B60 Dual uses PCIe 5.0 x8 and is a dual-slot card measuring 300 mm long. The RTX 5000 Ada uses PCIe 4.0 x16 and is dual-slot at 267 mm long. Intel has a faster bus standard but a longer physical card.

Architecture Differences

The Intel Arc Pro B60 Dual is built on the BMG-G21 chip using the Xe2-HPG architecture, part of the Battlemage Pro Series. It is fabricated on a 5 nm process at TSMC with 19,600 million transistors on a 272 mm² die. The transistor density measures 72.1 million per mm². The NVIDIA RTX 5000 Ada Generation uses the AD102 chip with Ada Lovelace architecture, also on a 5 nm TSMC process, but packs 76,300 million transistors on a 609 mm² die, yielding 125.3 million transistors per mm². The NVIDIA die is more than twice the size and holds nearly four times the transistors.

The shading resources differ dramatically. The Arc Pro B60 Dual has 2,560 shading units, 160 TMUs, and 80 ROPs. The RTX 5000 Ada has 12,800 shading units, 400 TMUs, and 176 ROPs. This 5x difference in shader count directly explains the FP32 throughput gap. The RTX 5000 Ada also includes 400 tensor cores, while the Arc Pro B60 Dual lists none.

Clock behavior shows a different design philosophy. Intel runs a base clock of 2000 MHz and boost of 2400 MHz. NVIDIA runs a lower base of 1155 MHz but boosts to 2550 MHz. The NVIDIA card achieves higher peak clocks despite a lower idle baseline, suggesting more aggressive boost behavior under load.

Memory architecture contrasts as well. Intel uses 24 GB GDDR6 at 2375 MHz (19 Gbps effective) over a 192-bit bus. NVIDIA uses 32 GB GDDR6 at 2250 MHz (18 Gbps effective) over a 256-bit bus. The wider bus gives NVIDIA higher bandwidth despite slightly slower memory clocks.

Power delivery differs substantially. The Arc Pro B60 Dual draws 400 W TDP with a suggested 800 W PSU. The RTX 5000 Ada draws 250 W TDP with a suggested 600 W PSU. Both use a single 16-pin connector and dual-slot cooling. Intel's higher power envelope buys a much smaller compute return per watt.

The bus interface also diverges. Intel uses PCIe 5.0 x8, while NVIDIA uses PCIe 4.0 x16. The x8 link on PCIe 5.0 provides similar bandwidth to x16 on PCIe 4.0, so neither has a practical bandwidth advantage in most workloads.

Display output standards differ. Intel offers 4x mini-DisplayPort 2.1, supporting higher refresh rates and bandwidth. NVIDIA offers 4x DisplayPort 1.4a, which is the older standard.

Production status for both is Active. Intel released in September 2025 with a launch MSRP of 1,199 USD. NVIDIA released in August 2023 and has no launch MSRP recorded. NVIDIA's predecessor is Workstation Ampere and its successor is Blackwell PRO W.

Head-to-Head Benchmarks

The head-to-head benchmark data shows only NVIDIA benchmarks recorded. The RTX 5000 Ada scores 175,286 in Geekbench OpenCL and 194,041 in Geekbench Vulkan. Its average benchmark score is 184,664. The Arc Pro B60 Dual has no recorded benchmark scores and no wins in the head-to-head comparison.

The percentile data confirms the gap. The RTX 5000 Ada sits in the 98th percentile among all GPUs. The Arc Pro B60 Dual sits in the 50th percentile. This places the NVIDIA card at the top tier of workstation graphics while Intel's card falls in the middle of the distribution.

The nearest rivals for the RTX 5000 Ada provide context. The NVIDIA A100 SXM4 80 GB scores 183,725, which is 0.5% below the RTX 5000 Ada. The A100 SXM4 40 GB scores 187,147, which is 1.3% higher. The RTX PRO 5000 Blackwell scores 182,109, which is 1.4% lower. The GeForce RTX 4090 D scores 178,050, which is 3.7% lower. The RTX 5000 Ada essentially trades blows with A100-class accelerators and beats the RTX 4090 D by nearly 4%.

The Vulkan score of 194,041 is notably higher than the OpenCL score of 175,286, a difference of about 10.7%. This suggests the Ada Lovelace architecture handles Vulkan workloads particularly well, possibly due to driver optimization or scheduling efficiency.

Without recorded benchmarks for the Arc Pro B60 Dual, direct comparison relies on architectural specifications. The FP32 throughput of 12.29 TFLOPS versus 65.28 TFLOPS indicates roughly a 5.3x gap in raw compute. Texture rate shows a similar pattern: 384.0 GTexel/s versus 1,020.0 GTexel/s, a 2.7x difference. Pixel rate is 192.0 GPixel/s versus 448.8 GPixel/s, a 2.3x gap.

The memory bandwidth difference is smaller proportionally. Intel's 456.0 GB/s trails NVIDIA's 576.0 GB/s by about 26%. This means memory-bound workloads will show a narrower gap than compute-bound workloads.

The RT core count difference is also substantial: 20 versus 100. Ray tracing performance will favor NVIDIA by a wide margin, though no specific ray tracing benchmarks exist in the database.

The Verdict

The data indicates the NVIDIA RTX 5000 Ada Generation is the dominant performer in nearly every measurable category. Its 98th percentile ranking versus Intel's 50th percentile places them in entirely different performance classes. The RTX 5000 Ada delivers 5.3x the FP32 throughput, 2.7x the texture rate, 2.3x the pixel rate, and 26% more memory bandwidth. It also carries 5x the RT cores and adds 400 tensor cores that Intel does not list at all.

The RTX 5000 Ada achieves this with lower power consumption. Its 250 W TDP is 150 W below Intel's 400 W TDP, and its 600 W suggested PSU is 200 W below Intel's 800 W recommendation. This efficiency advantage is significant for workstation builds where power delivery and cooling capacity matter.

The Arc Pro B60 Dual does have specific advantages. Its PCIe 5.0 x8 interface is a newer standard than PCIe 4.0 x16. Its DisplayPort 2.1 outputs support a newer display standard than DisplayPort 1.4a. Its 24 GB memory capacity, while smaller than 32 GB, is still substantial for many workloads.

The RTX 5000 Ada's benchmark scores place it alongside A100-class accelerators. The 0.5% delta against the A100 SXM4 80 GB and the 1.3% delta against the A100 SXM4 40 GB show it competes at data-center levels. Its 3.7% advantage over the RTX 4090 D confirms it outperforms a top-tier consumer card.

Users requiring maximum compute throughput, ray tracing capability, or tensor core acceleration should select the RTX 5000 Ada. Users prioritizing the newest display standard, PCIe 5.0 connectivity, or lower acquisition cost should consider the Arc Pro B60 Dual. The benchmark data, however, strongly favors NVIDIA for compute-intensive workloads.

Specification Differences

The two cards diverge across nearly every specification field.

Chip and architecture: Intel uses BMG-G21 with Xe2-HPG. NVIDIA uses AD102 with Ada Lovelace.

Process and die: Both use 5 nm TSMC. Intel has 19,600 million transistors on 272 mm². NVIDIA has 76,300 million on 609 mm². Transistor density is 72.1M/mm² for Intel and 125.3M/mm² for NVIDIA.

Clocks: Intel base 2000 MHz, boost 2400 MHz. NVIDIA base 1155 MHz, boost 2550 MHz. Memory clock is 2375 MHz (19 Gbps) for Intel and 2250 MHz (18 Gbps) for NVIDIA.

Memory: Intel has 24 GB GDDR6 on 192-bit with 456.0 GB/s. NVIDIA has 32 GB GDDR6 on 256-bit with 576.0 GB/s.

Compute units: Intel has 2,560 shading units, 160 TMUs, 80 ROPs, 20 RT cores. NVIDIA has 12,800 shading units, 400 TMUs, 176 ROPs, 100 RT cores, 400 tensor cores.

Rates: Intel pixel rate is 192.0 GPixel/s, texture rate 384.0 GTexel/s. NVIDIA pixel rate is 448.8 GPixel/s, texture rate 1,020.0 GTexel/s.

FP performance: Intel FP32 is 12.29 TFLOPS, FP16 is 24.58 TFLOPS (2:1). NVIDIA FP32 is 65.28 TFLOPS, FP16 is 65.28 TFLOPS (1:1).

Power: Intel TDP is 400 W with 800 W suggested PSU. NVIDIA TDP is 250 W with 600 W suggested PSU.

Physical: Intel is 300 mm long, 110 mm tall, 40 mm wide. NVIDIA is 267 mm long, 112 mm tall, width not recorded. Both are dual-slot.

Interface: Intel uses PCIe 5.0 x8. NVIDIA uses PCIe 4.0 x16.

Display outputs: Intel has 4x mini-DisplayPort 2.1. NVIDIA has 4x DisplayPort 1.4a.

APIs: Both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

Release: Intel released September 2025 with 1,199 USD launch MSRP. NVIDIA released August 2023 with no recorded MSRP.

Where Each One Wins

The RTX 5000 Ada Generation wins in compute-bound workloads. Its 65.28 TFLOPS FP32 throughput is 5.3x higher than Intel's 12.29 TFLOPS. This applies to scientific simulation, rendering, and any FP32-heavy application. Its FP16 performance of 65.28 TFLOPS at 1:1 ratio also exceeds Intel's 24.58 TFLOPS at 2:1 ratio, meaning NVIDIA maintains full-rate FP16 while Intel halves its rate.

The RTX 5000 Ada wins in ray tracing. Its 100 RT cores dwarf Intel's 20. Applications using RT acceleration will see a large performance gap.

The RTX 5000 Ada wins in AI and tensor workloads. Its 400 tensor cores provide dedicated hardware that Intel does not list. Machine learning inference and training acceleration are exclusive to NVIDIA.

The RTX 5000 Ada wins in memory capacity and bandwidth. Its 32 GB exceeds Intel's 24 GB, and its 576.0 GB/s exceeds Intel's 456.0 GB/s. Larger datasets and higher-resolution textures favor NVIDIA.

The RTX 5000 Ada wins in power efficiency. Its 250 W TDP delivers 65.28 TFLOPS, while Intel's 400 W TDP delivers 12.29 TFLOPS. The performance per watt ratio strongly favors NVIDIA.

The Arc Pro B60 Dual wins in display connectivity. Its 4x mini-DisplayPort 2.1 supports the newer standard, enabling higher resolutions and refresh rates over a single cable. NVIDIA's DisplayPort 1.4a is the previous generation.

The Arc Pro B60 Dual wins in bus standard. Its PCIe 5.0 x8 interface is newer than NVIDIA's PCIe 4.0 x16. For systems with PCIe 5.0 support, this provides equal bandwidth over fewer lanes.

The Arc Pro B60 Dual wins in physical footprint for length. At 300 mm it is longer than NVIDIA's 267 mm, but it is 2 mm shorter in height. Neither card has a clear physical advantage.

The Arc Pro B60 Dual wins on release timing. Its September 2025 release is newer than NVIDIA's August 2023 release, and its 1,199 USD launch MSRP is recorded in the database.

The Arc Pro B60 Dual wins in the 50th percentile placement only in the sense that it sits at the median of all GPUs. The RTX 5000 Ada at the 98th percentile leaves no performance category where Intel leads.