Intel Arc Pro B390 vs Intel Arc Pro B60 Comparison

Head-to-Head Benchmarks

The recorded data shows only one of these two Intel Arc Pro parts carries a benchmark profile. The Intel Arc Pro B60 has an average benchmark score of 3182, placing it in the 20th percentile of all GPUs in the database. The Intel Arc Pro B390 shows no recorded benchmark scores and sits at the 50th percentile with an average score of 0, meaning no direct apples-to-apples comparison can be drawn from the current measurements. What the data does allow is an examination of the B60's absolute performance profile and how it stacks against its nearest rivals.

The B60's strongest result comes from Passmark G3D, where it scores 14580. That figure towers over its own Passmark DirectX 11 result of 122, which highlights how differently synthetic suites weight various workloads. The Passmark GPU Compute score of 7029 is nearly half the G3D number, indicating compute-heavy tasks extract significantly less performance than rasterized graphics workloads. The 3DMark Steel Nomad DX12 result of 2646 is the only modern API benchmark recorded, and it shows a large gap from the legacy DirectX tests. Passmark DirectX 9 scores 179, DirectX 10 scores 61, and DirectX 12 scores 76, which suggests the card's legacy API performance does not scale linearly with its newer API capabilities.

Comparing the B60 to its nearest rivals, the delta percentages are tight but revealing. The B60's average score of 3182 sits 0.6% ahead of the NVIDIA Quadro P1000 (3163), making those two effectively neck-and-neck. Against the NVIDIA GeForce GT 640 (3210), the B60 trails by 0.9%. The GeForce 920M (3287) is 3.2% ahead, and the GeForce RTX 5080 SUPER (3075) is 3.5% behind. These margins are small, and the B60's 20th percentile ranking confirms it lands in the lower tier of the database's GPU distribution. The B390, lacking any benchmark entries, cannot be positioned within these rival clusters.

The absence of head-to-head benchmark data between the B390 and B60 means the only performance numbers available describe the B60 alone. The B390's 50th percentile is a positional marker, not a measured score. Any claim about which part wins in a specific workload would require recorded tests for both, and those tests do not exist in the database.

Architecture Differences

These two GPUs diverge at nearly every architectural layer. The Intel Arc Pro B390 uses the Panther Lake chip built on Intel's own 3 nm process, with the Xe3-LPG architecture. The Intel Arc Pro B60 uses the BMG-G21 chip fabricated by TSMC on a 5 nm process, with the Xe2-HPG architecture. The B60 is part of the Battlemage (Pro Series) generation, while the B390 belongs to the Arc Graphics-WM (Panther Lake) generation. The foundries differ: Intel for the B390, TSMC for the B60.

The B60 has published silicon details: 19,600 million transistors on a 272 mm² die, giving a transistor density of 72.1M per mm². The B390's transistor count and die size are listed as unknown, and no density figure exists. The B60's discrete design uses 24 GB of GDDR6 memory on a 192-bit bus, with 456.0 GB/s of bandwidth and memory clocked at 2375 MHz (19 Gbps effective). The B390 uses system shared memory, with bus width, memory type, and bandwidth all listed as system dependent. That is a fundamental split: one part carries dedicated VRAM, the other relies on the host system's memory pool.

The B390's base clock is 300 MHz with a boost of 2500 MHz. The B60 starts at 2000 MHz and boosts to 2400 MHz. The B390's lower base clock but higher boost suggests a power-adaptive design typical of integrated graphics, while the B60's higher floor and narrower boost range reflect a discrete card running sustained clocks. The B390 is an IGP with no power connectors and an 80 W TDP. The B60 is a dual-slot card with a single 8-pin connector, a 200 W TDP, and a suggested 550 W power supply.

Render output differs sharply. The B390 produces 60.00 GPixel/s and 120.0 GTexel/s, with 7.680 TFLOPS of FP32 and 15.36 TFLOPS of FP16 (2:1). The B60 delivers 192.0 GPixel/s, 384.0 GTexel/s, 12.29 TFLOPS of FP32, and 24.58 TFLOPS of FP16 (2:1). The B60 is 3.2x ahead in pixel rate, 3.2x in texture rate, 1.6x in FP32, and 1.6x in FP16. The B390's shading unit count is 1536 with 48 TMUs and 24 ROPs. The B60 has 2560 shading units, 160 TMUs, and 80 ROPs. Ray tracing cores: 12 for the B390, 20 for the B60.

Both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The B60 connects via PCIe 5.0 x8 and outputs four mini-DisplayPort 2.1 connectors. The B390's bus interface is IGP, and its display outputs are listed as portable device dependent. The B390 was released on 2026-01-26, the B60 on 2025-09-04. The B390 lists HD Graphics-WM as its predecessor; the B60 has no predecessor listed. The B60 has a launch MSRP of 499 USD.

The Verdict

The data supports a clear split in intended roles. The Intel Arc Pro B60 is a discrete workstation card with dedicated 24 GB GDDR6 memory, a 192-bit bus, and 456.0 GB/s of bandwidth. Its 200 W TDP and dual-slot cooler with a single 8-pin power connector mark it as a board meant for a desktop workstation chassis. Its benchmark results, while modest in percentile terms, are real and measurable. The B390 is an integrated graphics solution on Intel's own 3 nm process, sharing system memory, drawing 80 W, and using no power connectors. It targets portable or low-power devices where discrete graphics is not an option.

The B60's 20th percentile ranking places it below most discrete GPUs in the database, but its nearest rivals are all low-end or older parts: the Quadro P1000, GT 640, GeForce 920M, and RTX 5080 SUPER. The B390's 50th percentile is a placeholder, not a validated result, so no performance verdict can be issued for it. The B60's 3DMark Steel Nomad score of 2646 and Passmark G3D score of 14580 confirm it handles modern raster workloads, while its Passmark GPU Compute score of 7029 shows moderate compute throughput. The B390's specifications suggest it is a capable integrated option for compact systems, but the database contains no measurements to confirm how it performs.

For users requiring a discrete card with dedicated VRAM and a fixed power envelope, the B60 is the only one of these two with recorded performance data. For users needing an integrated GPU inside a Panther Lake platform, the B390 exists on paper with strong theoretical throughput for its class, but no empirical scores support any claim about its real-world speed.

Specification Differences

The two cards differ in every major specification category. The B390 uses a 3 nm Intel process with the Panther Lake chip and Xe3-LPG architecture. The B60 uses a 5 nm TSMC process with the BMG-G21 chip and Xe2-HPG architecture. The B60 has 19,600 million transistors on a 272 mm² die with a density of 72.1M per mm²; the B390's transistor count and die size are unknown. The B390's base clock is 300 MHz with a 2500 MHz boost; the B60's base is 2000 MHz with a 2400 MHz boost. Memory: the B390 uses system shared memory with system dependent bandwidth, while the B60 has 24 GB GDDR6 on a 192-bit bus with 456.0 GB/s.

The B390 has 1536 shading units, 48 TMUs, 24 ROPs, and 12 ray tracing cores. The B60 has 2560 shading units, 160 TMUs, 80 ROPs, and 20 ray tracing cores. Pixel rate is 60.00 GPixel/s for the B390 versus 192.0 GPixel/s for the B60. Texture rate is 120.0 GTexel/s versus 384.0 GTexel/s. FP32 is 7.680 TFLOPS versus 12.29 TFLOPS. FP16 is 15.36 TFLOPS versus 24.58 TFLOPS. TDP is 80 W versus 200 W. The B390 is an IGP with no power connectors; the B60 is dual-slot with one 8-pin connector and a suggested 550 W PSU. The B390 uses an IGP bus interface and portable device dependent display outputs; the B60 uses PCIe 5.0 x8 and four mini-DisplayPort 2.1 outputs. The B60 measures 167 mm by 69 mm by 40 mm; the B390 has no listed dimensions. Release dates: 2026-01-26 for the B390, 2025-09-04 for the B60.

FAQ

Q: Which GPU has more shading units?

A: The Intel Arc Pro B60 has 2560 shading units, while the Intel Arc Pro B390 has 1536 shading units.

Q: What memory configuration does each card use?

A: The B60 uses 24 GB of GDDR6 on a 192-bit bus with 456.0 GB/s bandwidth. The B390 uses system shared memory, with its bandwidth listed as system dependent.

Q: How do their FP32 compute performances compare?

A: The B60 delivers 12.29 TFLOPS of FP32, which is 1.6x the B390's 7.680 TFLOPS.

Q: What is the B60's average benchmark score and percentile?

A: The B60 has an average benchmark score of 3182 and sits in the 20th percentile of all GPUs. Its nearest rival, the NVIDIA Quadro P1000, scores 3163, putting the B60 0.6% ahead.

Q: Does the B390 have any recorded benchmark scores?

A: No. The B390 has an empty benchmark list and an average score of 0, with its 50th percentile serving only as a positional marker.

Q: Which card uses a smaller manufacturing process?

A: The B390 uses a 3 nm process from Intel, while the B60 uses a 5 nm process from TSMC.

Where Each One Wins

The Intel Arc Pro B60 wins in every measurable performance category. Its pixel rate of 192.0 GPixel/s is 3.2x the B390's 60.00 GPixel/s. Its texture rate of 384.0 GTexel/s is 3.2x the B390's 120.0 GTexel/s. FP32 throughput is 12.29 TFLOPS versus 7.680 TFLOPS, a 1.6x advantage. FP16 is 24.58 TFLOPS versus 15.36 TFLOPS, also 1.6x. The B60 has more shading units (2560 vs 1536), more TMUs (160 vs 48), more ROPs (80 vs 24), and more ray tracing cores (20 vs 12). It carries dedicated 24 GB GDDR6 memory with 456.0 GB/s bandwidth, whereas the B390 depends on system shared memory with bandwidth that varies by host platform.

The B390 wins in power efficiency and integration. Its 80 W TDP is 120 W lower than the B60's 200 W. It requires no power connectors and fits as an IGP, making it suitable for portable devices or compact platforms where a discrete dual-slot card with an 8-pin connector and 550 W suggested PSU cannot be accommodated. The B390's 3 nm process node is smaller than the B60's 5 nm node, and its higher boost clock of 2500 MHz exceeds the B60's 2400 MHz boost, though its 300 MHz base clock is far lower.

The B60 wins on connectivity and form factor. It offers four mini-DisplayPort 2.1 outputs and a PCIe 5.0 x8 interface, while the B390's display outputs are portable device dependent and its bus is IGP. The B60 also has published silicon data (19,600 million transistors, 272 mm² die, 72.1M per mm² density) and physical dimensions (167 mm by 69 mm by 40 mm), whereas the B390 lacks those details in the database.

The B60 is the only one with benchmark results. Its 3DMark Steel Nomad DX12 score of 2646, Passmark G3D score of 14580, and Passmark GPU Compute score of 7029 provide real-world reference points. The B390 has no such data, so any use-case claim for it rests solely on its architectural specifications: integrated design, low power draw, and system shared memory. The B60 is the choice for any workload requiring dedicated VRAM, sustained clocks, or discrete display outputs. The B390 is the choice for a system that cannot accommodate a discrete card at all.