Intel Arc B370

The Intel Arc B370 is an integrated graphics processor built on Intel's 3 nm process, using the Xe3-LPG architecture and the Panther Lake chip. It was released on January 26, 2026, and has a production status of Active. The part has 1280 shading units, 40 texture mapping units, 20 render output units, and 10 ray tracing cores. Its base clock is 300 MHz and its boost clock is 2400 MHz. Memory is system shared, with bandwidth listed as system dependent. The GPU is rated for a 25 W TDP and uses an IGP bus interface with no power connectors. It supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

How It Compares

The data pack lists no nearest rivals for the Arc B370, so there are no rival names, scores, or deltaPct values to cite. The only external position is the percentile field, which places the Arc B370 at the 50th percentile of all GPUs in the database. That is a median placement — half of the tracked GPUs rank above it and half below. The average benchmark score is 0, which indicates that no measured benchmark runs are recorded for this part, so the percentile may be a specification-based estimate rather than a result of tested performance. Without rival deltas, the comparison must rely on the Arc B370's own internal consistency. The 48 GPixel/s pixel rate, when divided by the 20 ROPs, is consistent with the 2400 MHz boost clock. The 96 GTexel/s texture rate, when divided by the 40 TMUs, also aligns with 2400 MHz. The 6.144 TFLOPS FP32 figure, when divided by the 1280 shading units, matches the same clock. These checks confirm that the listed rates are direct derivatives of the boost clock, giving a coherent theoretical throughput profile even in the absence of rival scores.

The absence of rival data means that the Arc B370 cannot be positioned against specific competing parts such as other integrated GPUs or low-end discrete cards. The 50th percentile is the sole ranking signal, and it places the part at the midpoint of the database's distribution. This is a neutral position — neither a standout nor a laggard — but it carries no information about which specific GPUs are nearby in the rankings. The internal consistency of the throughput rates at least verifies that the specifications are self-consistent, and the 2:1 FP16-to-FP32 ratio further confirms the part's compute profile is built around the same clock domain.

Who Should Consider It

The Arc B370 is an integrated part — the slot width is IGP, the bus interface is IGP, and the display outputs are portable device dependent. This makes it a candidate for thin-and-light laptops, 2-in-1 convertibles, and handheld gaming devices where a discrete GPU cannot fit. The 25 W TDP is low enough for such platforms, and the lack of power connectors means the GPU draws its power from the host socket. The memory configuration is entirely system shared, with bandwidth marked as system dependent, so the GPU's effective performance will hinge on the platform's memory speed and channel configuration. The data does not include game benchmarks, so resolution and settings recommendations cannot be derived from measured results. Instead, the compute rates provide a theoretical ceiling: 6.144 TFLOPS FP32 and 48 GPixel/s pixel rate are modest figures that point to light gaming, esports titles, and media acceleration rather than high-end 4K rendering. The FP16 rate of 12.29 TFLOPS, at a 2:1 ratio, offers headroom for compute workloads that can use reduced precision. API support for DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4 means the part can run modern graphics features, but the raw throughput will limit how much of that feature set can be exercised at high settings.

The display outputs being portable device dependent means the GPU's video output is tied to the host device's panel and ports, so it is not intended for external multi-monitor desktop setups. The system shared memory also means that the host's RAM capacity and speed directly affect the GPU's ability to hold textures and frame buffers. For users who prioritize battery life and low heat, the 25 W TDP is a strong fit. The 3 nm process helps keep the power envelope tight, and the 300 MHz base clock provides a low idle floor for power conservation when the GPU is not under load.

Power and Cooling

The Arc B370 is rated at 25 W TDP, a figure that defines its thermal envelope. It has no power connectors, and its bus interface is IGP, indicating that all power is delivered through the motherboard socket. The suggested PSU field is null, so the database offers no power supply recommendation; given the 25 W TDP, any PSU requirement would be dictated by the host system rather than this GPU. The slot width is IGP, meaning the part is not a discrete card and does not occupy a PCIe slot. Cooling is likewise integrated — the IGP form factor implies that the thermal solution is part of the device's main cooling assembly, shared with the CPU. The 3 nm process node from Intel's foundry contributes to the low power draw, and the wide clock range — 300 MHz base to 2400 MHz boost — allows the power management system to scale the GPU down to very low activity states or up to its peak when needed. The system dependent memory bandwidth means that power consumption for memory is not attributable to the GPU itself, as it uses the host's memory subsystem. The 300 MHz base clock provides a low idle floor, while the 2400 MHz boost clock allows the GPU to scale up for burst workloads. Because the data lists no suggested PSU, integrators must rely on the host platform's existing power delivery.

FAQ

Q: What is the boost clock of the Intel Arc B370?

A: The boost clock is 2400 MHz, with a base clock of 300 MHz.

Q: How much memory does the Arc B370 have?

A: The memory size, type, and bus width are all listed as "System Shared," and the bandwidth is "System Dependent." The GPU uses the host system's memory rather than dedicated VRAM.

Q: What graphics APIs does the Arc B370 support?

A: It supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

Q: What is the TDP of the Arc B370?

A: The TDP is 25 W. It has no power connectors and uses an IGP bus interface.

Q: When was the Arc B370 released?

A: The release date is January 26, 2026, and the production status is Active.

Q: What is the FP32 throughput of the Arc B370?

A: The FP32 throughput is 6.144 TFLOPS, with FP16 at 12.29 TFLOPS, a 2:1 ratio.

Benchmark Performance

The benchmark data for the Arc B370 is sparse: the benchmarks array is empty, the average benchmark score is 0, and the nearestRivals list is empty. Consequently, exact percentage deltas against rival products cannot be reported. What the data does provide is a set of theoretical throughput rates that can be analyzed for internal consistency and positioned against the database's overall percentile. The Arc B370's pixel rate is 48 GPixel/s, its texture rate is 96 GTexel/s, and its FP32 compute is 6.144 TFLOPS. The FP16 rate is 12.29 TFLOPS, which is exactly twice the FP32 rate, matching the 2:1 ratio stated in the data. The texture rate is exactly twice the pixel rate, a ratio that corresponds to the hardware configuration of 40 TMUs versus 20 ROPs — also a 2:1 ratio. The pixel rate divided by the 20 ROPs is consistent with the 2400 MHz boost clock, as is the texture rate divided by the 40 TMUs. The FP32 throughput divided by the 1280 shading units also aligns with the same clock. These internal consistencies indicate that all listed rates are derived from the same boost clock, giving a coherent theoretical performance profile.

The 10 ray tracing cores are present but have no dedicated throughput figure in the data, so their impact cannot be quantified. In the absence of measured scores, the 50th percentile ranking is the only comparative anchor. A median position in the database means the Arc B370 is neither a high-end nor a low-end part in the overall distribution, though the 25 W TDP and integrated form factor suggest that its real-world performance is bounded by the host system's memory and cooling. The system shared memory and system dependent bandwidth are the largest unknowns — they mean the GPU's performance will vary from one platform to another, making any single benchmark score inherently platform-specific. The 2:1 FP16-to-FP32 ratio is a common feature for GPUs that accelerate mixed-precision compute, and the 12.29 TFLOPS FP16 figure doubles the FP32 rate. The 10 ray tracing cores provide hardware acceleration for ray-traced effects, but without a dedicated RT throughput metric in the data, their performance cannot be compared to other parts. The API support for DirectX 12 Ultimate (12_2) includes features like hardware ray tracing and mesh shaders, which the RT cores can assist with. The 3 nm process is among the most advanced nodes in the database, which helps explain the low 25 W TDP. For a buyer or integrator, the data indicates a part that can handle light to moderate graphics workloads, with API support for modern titles, but without measured benchmarks, the exact frame rates and settings remain unspecified.