Intel Arc B370 vs NVIDIA L20 Comparison

Where Each One Wins

The Intel Arc B370 and NVIDIA L20 occupy entirely different segments of the graphics hardware landscape, and the recorded benchmark data reflects that separation clearly. The Intel Arc B370 is an integrated graphics processor built into the Panther Lake chip, designed for portable devices. Its single recorded benchmark, 3DMark Steel Nomad DX12, returns a score of 1184, which places it at the 5th percentile among all GPUs in the database. That percentile ranking means the vast majority of discrete graphics processors outperform it in this workload.

The NVIDIA L20, by contrast, is a server-class accelerator built on the Ada Lovelace architecture. Its benchmark results are dramatically higher. In Geekbench OpenCL, the L20 scores 274276, and in Geekbench Vulkan, it scores 228018. Its average benchmark score across these tests is 251147, placing it at the 99th percentile among all GPUs. The delta between the two products is not a matter of fine margins; it is a difference of several orders of magnitude in raw compute output.

The use-case split is therefore straightforward. The Intel Arc B370 wins in scenarios where power consumption and physical footprint are the primary constraints. Its thermal design power is 25 W, it uses no power connectors, and it is integrated directly into the processor, meaning it requires no expansion slot. The NVIDIA L20 demands a dual-slot chassis, a 16-pin power connector, a suggested 600 W power supply, and a PCIe 4.0 x16 slot. For thin-and-light portable devices, the Arc B370 is the only viable option between the two. For server workloads, scientific computing, and high-throughput rendering, the L20 is the only option that can deliver the required performance.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA L20 has an average benchmark score of 251147, while the Intel Arc B370 has an average benchmark score of 1184.

Q: How does the Intel Arc B370 compare to its nearest rivals?

A: The Arc B370 scores 1184 in 3DMark Steel Nomad DX12. It is 0.2% behind the ATI Mobility Radeon HD 5570, 0.5% behind the ATI Radeon HD 5770, 0.7% behind the AMD Radeon HD 7650M, and 1.4% ahead of the AMD FirePro M2000.

Q: How does the NVIDIA L20 compare to its nearest rivals?

A: The L20 is 11.6% ahead of the NVIDIA PG506-232, 14.2% ahead of the AMD Radeon PRO W7900D, 11.6% behind the NVIDIA L40, and 12.6% behind the NVIDIA RTX 6000 Ada Generation.

Q: What memory configurations do the two GPUs use?

A: The Intel Arc B370 uses system-shared memory with system-dependent bandwidth. The NVIDIA L20 has 48 GB of GDDR6 memory on a 384-bit bus with 864.0 GB/s of bandwidth.

Q: What is the process node for each GPU?

A: The Intel Arc B370 is built on a 3 nm process at Intel, while the NVIDIA L20 is built on a 5 nm process at TSMC.

Q: What is the thermal design power of each GPU?

A: The Intel Arc B370 has a TDP of 25 W, while the NVIDIA L20 has a TDP of 275 W.

Head-to-Head Benchmarks

The two products share no common benchmark tests in the database, so a direct apples-to-apples comparison requires examining their respective recorded scores and contextualizing them against their nearest rivals. The Intel Arc B370's only benchmark result is 1184 in 3DMark Steel Nomad DX12. That score places it just below the ATI Mobility Radeon HD 5570, which averages 1186, a delta of 0.2%. It also trails the ATI Radeon HD 5770 by 0.5% and the AMD Radeon HD 7650M by 0.7%. The only rival it beats is the AMD FirePro M2000, which averages 1168, putting the Arc B370 1.4% ahead.

The NVIDIA L20's Geekbench OpenCL score of 274276 and Geekbench Vulkan score of 228018 produce an average of 251147. Against its nearest rivals, the L20 sits comfortably above the NVIDIA PG506-232, which averages 225124, a lead of 11.6%. It also beats the AMD Radeon PRO W7900D, which averages 219827, by 14.2%. Above the L20 sit the NVIDIA L40 at 284111 and the NVIDIA RTX 6000 Ada Generation at 287237, putting the L20 11.6% and 12.6% behind those two respectively.

The compute throughput figures reinforce the gap. The Arc B370 delivers 6.144 TFLOPS of FP32 performance and 12.29 TFLOPS of FP16 performance at a 2:1 ratio. The L20 delivers 59.35 TFLOPS of FP32 and 59.35 TFLOPS of FP16 at a 1:1 ratio. The L20 also has significantly higher pixel and texture rates: 322.6 GPixel/s and 927.4 GTexel/s, versus 48.00 GPixel/s and 96.00 GTexel/s for the Arc B370. The L20's rasterization and texture throughput are roughly an order of magnitude higher across the board.

Clock speeds tell a similar story. The Arc B370 runs at a base clock of 300 MHz and a boost clock of 2400 MHz. The L20 runs at a base clock of 1440 MHz and a boost clock of 2520 MHz. While the boost clocks are somewhat close, the base clock difference is substantial, and the L20's much larger execution resource pool means every clock cycle accomplishes far more work.

Specification Differences

The two GPUs differ in nearly every measurable specification. The Intel Arc B370 uses a 3 nm process node fabricated by Intel, while the NVIDIA L20 uses a 5 nm process node fabricated by TSMC. The L20 has 76,300 million transistors on a 609 mm² die, with a transistor density of 125.3M per mm². The Arc B370's transistor count and die size are unknown.

Memory configurations are fundamentally different. The Arc B370 uses system-shared memory, meaning it has no dedicated VRAM, no dedicated memory bus width, and memory bandwidth that depends on the host system. The L20 has 48 GB of GDDR6 memory on a 384-bit bus, delivering 864.0 GB/s of bandwidth at an effective speed of 18 Gbps.

The compute resources differ by an order of magnitude. The Arc B370 has 1280 shading units, 40 texture mapping units, and 20 raster output units, along with 10 ray tracing cores. The L20 has 11776 shading units, 368 texture mapping units, and 128 raster output units, along with 92 ray tracing cores and 368 tensor cores. The Arc B370 has no listed tensor cores.

Power and physical requirements are also divergent. The Arc B370 has a TDP of 25 W, is integrated into the processor (IGP), requires no power connectors, and has no suggested PSU rating. The L20 has a TDP of 275 W, is a dual-slot card, requires one 16-pin power connector, and has a suggested PSU rating of 600 W. The L20 measures 267 mm in length and 111 mm in height. The Arc B370 has no listed dimensions because it is an integrated part.

The bus interface differs as well. The Arc B370 uses an integrated graphics processor interface, while the L20 uses PCIe 4.0 x16. Display outputs also differ: the Arc B370's display outputs are portable-device dependent, while the L20 provides 4x DisplayPort 1.4a outputs.

Architecture Differences

The Intel Arc B370 is built on the Xe3-LPG architecture and belongs to the Arc Graphics-M generation for Panther Lake. It uses the Panther Lake chip. The NVIDIA L20 uses the Ada Lovelace architecture with the AD102 chip and belongs to the Server Ada (Lxx) generation.

The process technology separates the two products by foundry and node. Intel manufactures the Arc B370 on its own 3 nm process. TSMC manufactures the L20 on a 5 nm process. The L20's transistor count of 76,300 million and die size of 609 mm² are recorded, while the Arc B370's are not disclosed.

The ray tracing and tensor core configurations highlight their different design goals. The Arc B370 has 10 ray tracing cores and no tensor cores, suited for basic graphics acceleration in an integrated context. The L20 has 92 ray tracing cores and 368 tensor cores, reflecting its server positioning for AI and compute-heavy workloads.

API support is identical between the two: both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The FP16 ratio differs, with the Arc B370 running at 2:1 and the L20 at 1:1, meaning the L20 maintains full FP16 throughput without the halving penalty.

Release timing also separates them. The Arc B370 was released on January 26, 2026, while the L20 released on November 15, 2023. The L20's predecessor is Server Ampere and its successor is Server Hopper. The Arc B370 has no recorded predecessor or successor. Both are listed as Active in production status.

The Verdict

The recorded data indicates two products built for entirely different environments with no meaningful overlap in target workload. The Intel Arc B370 is an integrated GPU with a 25 W TDP, system-shared memory, and a 3DMark Steel Nomad DX12 score of 1184. It ranks at the 5th percentile among all GPUs, placing it among the lower-performing graphics solutions in the database. Its nearest rivals are all older mobile and desktop parts from the ATI and AMD families, and it trades fractions of a percent with them.

The NVIDIA L20 is a server accelerator with 48 GB of GDDR6 memory, 864.0 GB/s of bandwidth, 59.35 TFLOPS of FP32 compute, and an average benchmark score of 251147. It ranks at the 99th percentile among all GPUs, placing it among the highest-performing accelerators in the database. Its nearest rivals are professional and server-class parts like the NVIDIA PG506-232, AMD Radeon PRO W7900D, NVIDIA L40, and NVIDIA RTX 6000 Ada Generation.

Users of portable devices with Panther Lake processors will find the Arc B370 integrated directly into the chip, requiring no additional power delivery or cooling beyond what the host system already provides. Server operators and compute-intensive workloads will require the L20's dedicated memory, tensor cores, and high FP32 throughput. The data does not suggest any scenario where these two products compete for the same purchase decision. The choice is determined entirely by form factor, power envelope, and workload requirements.