Intel Arc Pro B60 vs NVIDIA GeForce RTX 3060 Ti GDDR6X Comparison

Intel
GPU

Intel Arc Pro B60

CORE STATE BMG-G21
VRAM 24 GB
CLOCK SPEED 2400 MHz
TDP 200 W
BUS WIDTH 192 bit
ARCHITECTURE Xe2-HPG
nm
PROCESS 5 nm
LAUNCH DATE 2025
VS
NVIDIA
GEFORCE

GeForce RTX 3060 Ti GDDR6X

CORE STATE GA104
VRAM 8 GB
CLOCK SPEED 1665 MHz
TDP 225 W
BUS WIDTH 256 bit
ARCHITECTURE Ampere
nm
PROCESS 8 nm
LAUNCH DATE 2022

PERFORMANCE BENCHMARKS

3dmark_3dmark_steel_nomad_dx12
2,646
2,795
passmark_directx_10
61
N/A
passmark_directx_11
122
N/A
passmark_directx_12
76
N/A
passmark_directx_9
179
N/A
passmark_g2d
763
N/A
passmark_g3d
14,580
N/A
passmark_gpu_compute
7,029
N/A

Analysis: Intel Arc Pro B60 vs NVIDIA GeForce RTX 3060 Ti GDDR6X

Intel Arc Pro B60 and NVIDIA GeForce RTX 3060 Ti GDDR6X occupy different corners of the GPU landscape, yet their benchmark scores place them in a surprisingly narrow band. The data shows the RTX 3060 Ti GDDR6X winning the sole direct head-to-head comparison, but the Arc Pro B60 counters with a substantially larger memory pool and a more modern feature set. This analysis breaks down what the numbers reveal about these two very different designs.

Head-to-Head Benchmarks

The only shared benchmark in the data is 3DMark Steel Nomad (DX12), and it produces a clear, if modest, victory for the NVIDIA GeForce RTX 3060 Ti GDDR6X. The RTX 3060 Ti scores 2795, while the Intel Arc Pro B60 trails with 2646. That is a delta of -5.3% for the Intel part, meaning NVIDIA’s card is roughly 5.3% faster in this specific test. While not a landslide, the result is consistent with the RTX 3060 Ti’s higher raw FP32 throughput — 16.20 TFLOPS versus 12.29 TFLOPS — and its larger 256-bit memory bus delivering 608.3 GB/s of bandwidth.

However, the broader benchmark picture is less straightforward. The Intel Arc Pro B60 has a much richer benchmark profile, with eight recorded scores spanning older DirectX versions and compute workloads. Its Passmark G3D score is 14580, and its Passmark GPU Compute score is 7029. The NVIDIA card has only the single Steel Nomad result in this data set, so direct apples-to-apples comparisons across other tests are impossible from the FACT PACK. What the data does show is that the Arc Pro B60’s average benchmark score of 3182 sits just 0.6% above the NVIDIA Quadro P1000 (3163) and 3.5% above the NVIDIA GeForce RTX 5080 SUPER (3075). In contrast, the RTX 3060 Ti’s average score of 2795 places it 0.4% above the NVIDIA GeForce 820M (2784) and 1.6% above the Intel UHD Graphics 610 (2752).

The percentile rankings reinforce the idea that neither card is a top-tier performer. The Arc Pro B60 sits in the 20th percentile of all GPUs, while the RTX 3060 Ti sits just below at the 18th percentile. These are close, mid-range positions. The win count is 1-0 in favor of NVIDIA in direct head-to-head tests, but that single test does not capture the Intel card’s strengths in other areas, particularly its 24 GB memory capacity and its 456.0 GB/s bandwidth, which, while lower than NVIDIA’s, is paired with a much larger framebuffer.

FAQ

Q: Which GPU wins the only direct benchmark comparison?

A: The NVIDIA GeForce RTX 3060 Ti GDDR6X wins the 3DMark Steel Nomad (DX12) test with a score of 2795 versus the Intel Arc Pro B60’s 2646, a 5.3% advantage for NVIDIA.

Q: How do the two compare in terms of memory capacity?

A: The Intel Arc Pro B60 has 24 GB of GDDR6 memory, which is three times the 8 GB of GDDR6X found on the NVIDIA card. This is a significant difference, especially for workloads that require large datasets to reside in VRAM.

Q: What is the difference in their memory bandwidth figures?

A: The NVIDIA card has higher memory bandwidth at 608.3 GB/s, whereas the Intel card offers 456.0 GB/s. Despite the lower bandwidth, the Intel card’s larger memory pool may compensate in capacity-bound scenarios.

Q: Are there differences in their FP32 compute performance?

A: Yes, the NVIDIA GeForce RTX 3060 Ti GDDR6X has a higher FP32 rating of 16.20 TFLOPS, compared to 12.29 TFLOPS for the Intel Arc Pro B60. This aligns with NVIDIA’s win in the 3DMark test.

Q: How do their overall benchmark averages compare?

A: The Intel Arc Pro B60 has a higher average benchmark score of 3182, whereas the NVIDIA card’s average is 2795. However, this comparison is limited because the Intel card has eight benchmark results, while the NVIDIA card has only one.

Q: What are the production statuses of the two GPUs?

A: The Intel Arc Pro B60 is listed as "Active," while the NVIDIA GeForce RTX 3060 Ti GDDR6X is listed as "End-of-life." This indicates the Intel product is currently in production, whereas the NVIDIA product has been discontinued.

Architecture Differences

The two GPUs are built on fundamentally different architectures and process nodes. Intel’s Arc Pro B60 uses the Xe2-HPG architecture with the BMG-G21 chip, fabricated on a 5 nm process at TSMC. NVIDIA’s RTX 3060 Ti GDDR6X uses the older Ampere architecture with the GA104 chip, built on an 8 nm process at Samsung. The process node difference is significant: the Intel chip packs 19,600 million transistors into a 272 mm² die, yielding a transistor density of 72.1M per mm². The NVIDIA chip, despite having fewer transistors at 17,400 million, has a larger die size of 392 mm², resulting in a much lower density of 44.4M per mm². This indicates Intel’s design is more area-efficient.

The memory subsystems diverge sharply. The Intel card uses 24 GB of GDDR6 on a 192-bit bus, while the NVIDIA card uses 8 GB of GDDR6X on a 256-bit bus. The GDDR6X on NVIDIA is faster, providing 608.3 GB/s of bandwidth versus 456.0 GB/s for Intel. The difference in bus width and memory type explains this trade-off. The Intel card’s triple capacity could be a decisive factor for professional workloads like large AI model inference or high-resolution texture rendering, where memory capacity is the primary bottleneck.

Compute resources also differ. NVIDIA’s card has 4864 shading units, 152 TMUs, 38 RT cores, and 152 tensor cores. Intel’s card has 2560 shading units, 160 TMUs, and 20 RT cores. Intel’s higher TMU count (160 vs 152) is notable, but NVIDIA’s shading unit advantage is massive. The FP32 throughput difference (16.20 vs 12.29 TFLOPS) reflects this, though Intel’s FP16 performance is rated at 24.58 TFLOPS (2:1 ratio), while NVIDIA’s is 16.20 TFLOPS (1:1). This suggests Intel may have an advantage in workloads that can leverage FP16 with a 2:1 throughput ratio.

Other architectural features include PCIe interface support. Intel uses PCIe 5.0 x8, while NVIDIA uses PCIe 4.0 x16. While the older NVIDIA standard has more lanes, the newer PCIe 5.0 standard on Intel offers higher per-lane bandwidth, potentially evening out the interface throughput. Display outputs also differ, with Intel offering four mini-DisplayPort 2.1 connections, while NVIDIA provides one HDMI 2.1 and three DisplayPort 1.4a outputs. Both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4, so no API feature-level differentiation exists in the data.

The Verdict

The data does not present a single outright winner; it presents a choice between two distinct trade-offs. For raw rasterization performance in a single tested workload, the NVIDIA GeForce RTX 3060 Ti GDDR6X is faster, delivering a 5.3% higher score in 3DMark Steel Nomad. Its higher FP32 throughput and memory bandwidth make it the stronger choice for applications that are compute-bound or bandwidth-sensitive.

For workloads that demand large memory capacity, the Intel Arc Pro B60 is the clear option. Its 24 GB VRAM is three times the NVIDIA card’s 8 GB. This is not a minor detail; it is a defining characteristic. The Intel card also has a higher average benchmark score (3182 vs 2795) and a newer production status (Active vs End-of-life). The process node advantage (5 nm vs 8 nm) and higher transistor density suggest a more modern, efficient design.

The choice depends on the primary use case. If the task is gaming or general-purpose rendering where the 3DMark result is representative, the NVIDIA card offers better performance. If the task is large-scale data processing, AI inferencing, or any workload that exceeds 8 GB of VRAM, the Intel card is the only viable option from the data presented. The RTX 3060 Ti’s end-of-life status also makes the Intel card a more forward-looking purchase for new system builds, despite its lower peak performance in the single benchmark.

Specification Differences

| Specification | Intel Arc Pro B60 | NVIDIA GeForce RTX 3060 Ti GDDR6X |

| :--- | :--- | :--- |

| Architecture | Xe2-HPG (Battlemage) | Ampere (GeForce 30) |

| Process Node | 5 nm (TSMC) | 8 nm (Samsung) |

| Transistors | 19,600 million | 17,400 million |

| Die Size | 272 mm² | 392 mm² |

| Transistor Density | 72.1M / mm² | 44.4M / mm² |

| Base Clock | 2000 MHz | 1410 MHz |

| Boost Clock | 2400 MHz | 1665 MHz |

| Memory Size | 24 GB GDDR6 | 8 GB GDDR6X |

| Memory Bus Width | 192 bit | 256 bit |

| Memory Bandwidth | 456.0 GB/s | 608.3 GB/s |

| Shading Units | 2560 | 4864 |

| TMUs | 160 | 152 |

| Pixel Rate | 192.0 GPixel/s | 133.2 GPixel/s |

| Texture Rate | 384.0 GTexel/s | 253.1 GTexel/s |

| FP32 Compute | 12.29 TFLOPS | 16.20 TFLOPS |

| FP16 Compute | 24.58 TFLOPS (2:1) | 16.20 TFLOPS (1:1) |

| RT Cores | 20 | 38 |

| TDP | 200 W | 225 W |

| Power Connector | 1x 8-pin | 1x 12-pin |

| Bus Interface | PCIe 5.0 x8 | PCIe 4.0 x16 |

| Display Outputs | 4x mini-DisplayPort 2.1 | 1x HDMI 2.1, 3x DisplayPort 1.4a |

| Production Status | Active | End-of-life |

| Release Date | 2025-09-04 | 2022-10-18 |

DETAILED SPECIFICATIONS

SPECIFICATION
Pro B60
RTX 3060 Ti GDDR6X
Core Specs
Shading Units
2,560
4,864 +90.0%
Shaders
2,560
4,864 +90.0%
TMUs
160
152 -5.0%
ROPs
80
80 0.0%
SM Count
—
38
Execution Units
20
—
Clocks
Base Clock
2000 MHz
1410 MHz
Boost Clock
2400 MHz
1665 MHz
Memory Clock
2375 MHz 19 Gbps effective
1188 MHz 19 Gbps effective
Memory
Memory Size
24 GB
8 GB
VRAM (MB)
24,576
8,192 -66.7%
Memory Type
GDDR6
GDDR6X
Memory Bus
192 bit
256 bit
Bandwidth
456.0 GB/s
608.3 GB/s
Cache
L1 Cache
—
128 KB (per SM)
L2 Cache
10 MB
4 MB
Performance
Pixel Rate
192.0 GPixel/s
133.2 GPixel/s
Texture Rate
384.0 GTexel/s
253.1 GTexel/s
FP32 (TFLOPS)
12.29 TFLOPS
16.20 TFLOPS
FP64 (TFLOPS)
3.072 TFLOPS (1:4)
253.1 GFLOPS (1:64)
FP16 (TFLOPS)
24.58 TFLOPS (2:1)
16.20 TFLOPS (1:1)
AI/RT
RT Cores
20
38 +90.0%
Tensor Cores
—
152
XMX Cores
160
—
Power
TDP
200 W
225 W
TDP (W)
200
225 +12.5%
Suggested PSU
550 W
550 W
Power Connectors
1x 8-pin
1x 12-pin
Architecture
Architecture
Xe2-HPG
Ampere
GPU Name
BMG-G21
GA104
Generation
Battlemage (Pro Series)
GeForce 30
Process Size
5 nm
8 nm
Transistors
19,600 million
17,400 million
Die Size
272 mm²
392 mm²
Foundry
TSMC
Samsung
Density
72.1M / mm²
44.4M / mm²
API Support
DirectX
12 Ultimate (12_2)
12 Ultimate (12_2)
OpenGL
4.6
4.6
Vulkan
1.4
1.4
OpenCL
3.0
3.0
CUDA
—
8.6
Shader Model
6.6
6.8
Physical
Slot Width
Dual-slot
Dual-slot
Length
167 mm 6.6 inches
242 mm 9.5 inches
Height
69 mm 2.7 inches
112 mm 4.4 inches
Outputs
4x mini-DisplayPort 2.1
1x HDMI 2.13x DisplayPort 1.4a
Bus Interface
PCIe 5.0 x8
PCIe 4.0 x16
Other
Launch Price
499 USD
—
Production
Active
End-of-life
Predecessor
—
GeForce 20
Successor
—
GeForce 40
View Arc Pro B60 Details View GeForce RTX 3060 Ti GDDR6X Details