GPU Comparison

# Intel Arc Pro B60 vs NVIDIA GeForce 825M

The Intel Arc Pro B60 and NVIDIA GeForce 825M occupy entirely different tiers of the GPU landscape, separated by more than a decade of architectural progress and a 167-watt power envelope gap. The Arc Pro B60, built on TSMC's 5 nm process with 19,600 million transistors, delivers 12.29 TFLOPS of FP32 compute and 456.0 GB/s of memory bandwidth, while the GeForce 825M, a 28 nm Kepler part with 1,020 million transistors, manages 722.7 GFLOPS and 14.40 GB/s. Benchmark data confirms this chasm: the Arc Pro B60 scores 14,580 in Passmark G3D and 7,029 in GPU compute, whereas the 825M's only recorded benchmark is a Geekbench OpenCL score of 3,694. The Arc Pro B60 sits at the 20th percentile of all GPUs, with an average benchmark score of 3,182, while the 825M ranks at the 22nd percentile with an average score of 3,694, a paradox explained by the fact that these averages draw from different benchmark suites with different scaling.

The Verdict

The data presents a straightforward recommendation for nearly every use case: the Intel Arc Pro B60 is the superior choice for anyone needing modern graphics features, substantial memory capacity, or professional-grade compute performance. Its 24 GB of GDDR6 memory dwarfs the 825M's 1 GB of DDR3, and its 2,560 shading units versus 384 means the Arc Pro B60 has roughly 6.7 times the raw shader throughput. The Arc Pro B60 supports DirectX 12 Ultimate (12_2), Vulkan 1.4, and OpenGL 4.6, while the 825M is limited to DirectX 12 (11_0) and Vulkan 1.2.175, a feature gap that matters for any modern workload.

However, the GeForce 825M is not without a niche. Its 33 W TDP and IGP (integrated graphics processor) form factor make it suitable for compact, low-power systems where the Arc Pro B60's 200 W TDP and dual-slot cooler are non-starters. The 825M requires no power connectors and has no suggested PSU rating, whereas the Arc Pro B60 needs a 1x 8-pin connector and a 550 W power supply. For legacy applications that predate modern APIs, the 825M's Kepler architecture may offer compatibility advantages, though the benchmark data does not explicitly confirm this.

Purchasers seeking a professional workstation GPU with future-proofing should choose the Intel Arc Pro B60 without hesitation. Its 4x mini-DisplayPort 2.1 outputs and PCIe 5.0 x8 interface provide modern connectivity that the 825M's "Portable Device Dependent" display outputs cannot match. The 825M, being end-of-life since its 2014 release, offers no upgrade path and no support for ray tracing, while the Arc Pro B60 includes 20 dedicated RT cores.

Architecture Differences

The architectural gap between these two GPUs spans five process nodes and two fundamentally different design philosophies. The Intel Arc Pro B60 uses the BMG-G21 chip with Xe2-HPG architecture, part of the Battlemage (Pro Series) generation, fabricated on TSMC's 5 nm process. This yields a transistor density of 72.1 million per square millimeter across a 272 mm² die, totaling 19,600 million transistors. In contrast, the NVIDIA GeForce 825M uses the GK208 chip with Kepler 2.0 architecture from the GeForce 800M generation, built on TSMC's 28 nm process. Its density is just 11.7 million transistors per square millimeter across an 87 mm² die, totaling 1,020 million transistors, roughly 5% of the Arc Pro B60's transistor count.

Clock speeds tell a similar story. The Arc Pro B60 operates at a 2000 MHz base clock and 2400 MHz boost, with memory running at 2375 MHz (19 Gbps effective). The 825M's base clock is 850 MHz with a 941 MHz boost, and its memory runs at 900 MHz (1800 Mbps effective). Despite the 825M's lower clocks, the Arc Pro B60's advantages are compounded by its wider memory bus (192-bit versus 64-bit) and newer GDDR6 memory versus DDR3, resulting in 456.0 GB/s versus 14.40 GB/s bandwidth, a 31.7x difference.

Compute resources further differentiate the two. The Arc Pro B60 features 2,560 shading units, 160 texture mapping units, 80 ROPs, and 20 RT cores, delivering 192.0 GPixel/s pixel fill rate and 384.0 GTexel/s texture rate. The 825M has 384 shading units, 32 TMUs, and 8 ROPs, achieving 7.528 GPixel/s and 30.11 GTexel/s. The Arc Pro B60's FP32 throughput of 12.29 TFLOPS is 17 times the 825M's 722.7 GFLOPS, and the Arc Pro B60 also supports FP16 at 24.58 TFLOPS (2:1 ratio), a capability the 825M lacks entirely.

The Arc Pro B60's API support is notably more modern: DirectX 12 Ultimate (12_2), Vulkan 1.4, and OpenGL 4.6. The 825M supports DirectX 12 (11_0), Vulkan 1.2.175, and OpenGL 4.6. The Arc Pro B60's PCIe 5.0 x8 interface doubles the bandwidth of the 825M's PCIe 3.0 x8, and its 167 mm length with dual-slot cooling contrasts with the 825M's IGP form factor.

Where Each One Wins

The Intel Arc Pro B60 wins decisively in every compute-heavy and graphics-intensive scenario. Its Passmark G3D score of 14,580 versus the 825M's absence from that benchmark (only a Geekbench OpenCL score of 3,694 is recorded) indicates that the Arc Pro B60 is designed for demanding 3D workloads, content creation, and professional visualization. The Arc Pro B60's 24 GB memory capacity handles large datasets and high-resolution textures that would exhaust the 825M's 1 GB frame buffer instantly. Its 20 RT cores enable hardware-accelerated ray tracing, a feature entirely absent from the 825M.

The GeForce 825M's advantages are confined to power-constrained and legacy scenarios. Its 33 W TDP means it can be integrated into thin-and-light laptops or small form factor systems where the Arc Pro B60's 200 W TDP is impractical. The 825M's IGP form factor eliminates the need for discrete power connectors, making it suitable for systems without PCIe power delivery. Its predecessor relationship to the GeForce 700M and successor to the GeForce 900M indicates it was part of a long-lived mobile lineup, suggesting driver maturity for older operating systems.

For professional workloads, the Arc Pro B60's 4x mini-DisplayPort 2.1 outputs support multi-monitor configurations with modern display standards. The 825M's "Portable Device Dependent" outputs tie it to specific laptop designs, offering no flexibility. The Arc Pro B60's production status is "Active," while the 825M is "End-of-life," meaning the Intel part will receive ongoing support and the NVIDIA part will not.

FAQ

Q: Which GPU has higher raw compute performance?

A: The Intel Arc Pro B60 delivers 12.29 TFLOPS FP32 performance, which is approximately 17 times the GeForce 825M's 722.7 GFLOPS. The Arc Pro B60 also supports FP16 at 24.58 TFLOPS, a capability the 825M does not offer.

Q: How do their memory subsystems compare?

A: The Arc Pro B60 has 24 GB of GDDR6 memory on a 192-bit bus with 456.0 GB/s bandwidth. The GeForce 825M has 1 GB of DDR3 memory on a 64-bit bus with 14.40 GB/s bandwidth. This represents a 31.7x difference in memory bandwidth.

Q: What are the power requirements for each card?

A: The Arc Pro B60 has a 200 W TDP, requires a 1x 8-pin power connector, and suggests a 550 W power supply. The GeForce 825M has a 33 W TDP, requires no power connectors, and has no suggested PSU rating, making it suitable for low-power systems.

Q: Which GPU supports newer graphics APIs?

A: The Arc Pro B60 supports DirectX 12 Ultimate (12_2), Vulkan 1.4, and OpenGL 4.6. The GeForce 825M supports DirectX 12 (11_0), Vulkan 1.2.175, and OpenGL 4.6. The Arc Pro B60's API set is more modern and feature-complete.

Q: What is the production status of each GPU?

A: The Intel Arc Pro B60 is listed as "Active" production, released on 2025-09-04 with a launch MSRP of 499 USD. The NVIDIA GeForce 825M is "End-of-life," released on 2014-01-26, with the GeForce 700M as predecessor and GeForce 900M as successor.

Q: How do their benchmark scores compare?

A: The Arc Pro B60 scores 14,580 in Passmark G3D, 7,029 in Passmark GPU Compute, and 2,646 in 3DMark Steel Nomad DX12. The GeForce 825M's only recorded benchmark is 3,694 in Geekbench OpenCL. The Arc Pro B60's average benchmark score is 3,182 versus 3,694 for the 825M, though these averages use different benchmark suites.

Head-to-Head Benchmarks

Direct head-to-head benchmark comparisons are unavailable in the data, as the two GPUs were tested with entirely different benchmark suites. However, the available scores reveal the performance hierarchy clearly. The Intel Arc Pro B60 achieves 14,580 in Passmark G3D, a score that places it in the 20th percentile of all GPUs. Its nearest rivals include the NVIDIA Quadro P1000 (average score 3,163, delta 0.6%), NVIDIA GeForce GT 640 (3,210, delta -0.9%), NVIDIA GeForce 920M (3,287, delta -3.2%), and NVIDIA GeForce RTX 5080 SUPER (3,075, delta 3.5%). These deltas show the Arc Pro B60 performing within 3.5% of its closest competitors, indicating it occupies a highly competitive segment of the market.

The Arc Pro B60's Passmark DirectX scores show version-dependent performance: 179 in DirectX 9, 122 in DirectX 11, 76 in DirectX 12, and 61 in DirectX 10. Its Passmark G2D score is 763, and its GPU compute score is 7,029. The 3DMark Steel Nomad DX12 score of 2,646 provides a modern, DirectX 12-specific measure of gaming performance.

The NVIDIA GeForce 825M's single benchmark entry, Geekbench OpenCL at 3,694, places it in the 22nd percentile of all GPUs. Its nearest rivals are NVIDIA GeForce GT 740M (3,717, delta -0.6%), NVIDIA Quadro 3000M (3,718, delta -0.6%), AMD Radeon HD 6770 (3,649, delta 1.2%), and NVIDIA GeForce GT 635M (3,740, delta -1.2%). The 825M's performance is within 1.2% of these competitors, showing that it performed at parity with mid-range mobile GPUs of its era.

The critical observation is that the Arc Pro B60's Passmark G3D score of 14,580 is roughly 3.9 times the 825M's Geekbench OpenCL score of 3,694, even though these tests measure different aspects of performance. In GPU compute, the Arc Pro B60's 7,029 score is 1.9 times the 825M's OpenCL result. The Arc Pro B60's 3DMark Steel Nomad score of 2,646, a modern DirectX 12 benchmark, has no equivalent for the 825M, which predates the test's requirements.

The Arc Pro B60's 24.58 TFLOPS FP16 capability and 12.29 TFLOPS FP32 throughput dwarf the 825M's 722.7 GFLOPS FP32. Even accounting for architectural efficiency differences between Xe2-HPG and Kepler 2.0, the raw compute advantage is overwhelming. The Arc Pro B60's 456.0 GB/s memory bandwidth versus 14.40 GB/s represents a 31.7x advantage, which is critical for texture-heavy workloads, large datasets, and high-resolution rendering. The 825M's 8 ROPs versus 80 ROPs on the Arc Pro B60 means a 10x difference in pixel fill rate (192.0 GPixel/s versus 7.528 GPixel/s), directly impacting rasterization performance at high resolutions.