Intel Arc B390 vs Intel Arc B770 Comparison

Where Each One Wins

The Intel Arc B390 and Intel Arc B770 occupy entirely different positions in the performance spectrum, and the recorded data reflects a clear separation of roles rather than a direct competition.

The Intel Arc B390 is an integrated graphics processor built on the Panther Lake chip with the Xe3-LPG architecture. It operates as an IGP, meaning it shares system memory and depends on the host platform for its memory bandwidth. Its benchmark presence is defined by a single recorded result: a 3DMark Steel Nomad DX12 score of 1482. That score places it in the 9th percentile of all GPUs in the database, which indicates that it sits near the lower end of the performance range. Its nearest rivals, according to the database, are all older NVIDIA discrete parts: the GeForce GT 520MX at 1463, the GeForce 800M at 1460, the GeForce GT 625 OEM at 1446, and the GeForce GT 710 at 1443. The B390 leads each of these by a narrow margin, with deltas of 1.3%, 1.5%, 2.5%, and 2.7% respectively. These are small but consistent advantages, and they suggest that the B390 delivers entry-level performance that can edge out some of the most modest dedicated GPUs from prior generations.

The Intel Arc B770, in contrast, is a discrete add-in board based on the BMG-G31 chip with the Xe2-HPG architecture, belonging to the Battlemage generation in the Arc 7 series. It carries 16 GB of GDDR6 memory on a 256-bit bus, with a recorded bandwidth of 512.0 GB/s. Its compute resources are substantially larger: 4096 shading units, 256 texture mapping units, 128 raster operation units, and 32 ray tracing cores. The B770 has no recorded benchmark scores in the database, so its performance cannot be directly quantified against the B390 or any other GPU. However, its specifications alone indicate a completely different performance class. The B770 is rated for 19.66 TFLOPS of FP32 compute and 39.32 TFLOPS of FP16 compute, while the B390 is rated for 7.680 TFLOPS FP32 and 15.36 TFLOPS FP16. The B770's pixel rate of 307.2 GPixel/s and texture rate of 614.4 GTexel/s dwarf the B390's 60.00 GPixel/s and 120.0 GTexel/s.

The B770 is listed in the 50th percentile of all GPUs, which is the median position. That placement, combined with its lack of benchmark entries, suggests that the database treats it as a mid-pack performer by classification, though not yet validated with actual test scores. The B390, by contrast, is at the 9th percentile, which is firmly in the low end. The wins are therefore not a matter of one card beating the other in a series of tests; there are no head-to-head benchmark entries between these two products. Instead, the distinction is architectural and operational. The B390 wins in the category of integration, requiring no power connectors, drawing 80 W, and fitting into an IGP slot. The B770 wins in every raw capability metric recorded in the database, from memory capacity to compute throughput, but it does so at the cost of a 225 W TDP, a dual-slot footprint, and a 550 W suggested power supply.

The Verdict

The data does not present a single winner because the two products are not aimed at the same use case. The Intel Arc B390 is an integrated solution. Its 80 W TDP, IGP form factor, and lack of power connectors mean it is designed to be embedded into a portable device or a compact platform where discrete graphics are not an option. Its benchmark score of 1482 in 3DMark Steel Nomad DX12, while low in absolute terms, is enough to surpass the GeForce GT 710 by 2.7%, the GeForce GT 625 OEM by 2.5%, the GeForce 800M by 1.5%, and the GeForce GT 520MX by 1.3%. For a user constrained to integrated graphics, the B390 provides a measurable, if modest, improvement over some of the weakest discrete GPUs still present in the database.

The Intel Arc B770 is a discrete card for a completely different workload. With 16 GB of GDDR6, a 256-bit bus, 512.0 GB/s of bandwidth, and 4096 shading units, it is built for high-resolution rendering and compute-heavy tasks. The lack of recorded benchmark scores means the database cannot confirm how it performs in practice, but its specification sheet places it far above the B390 in every measurable compute category. The FP32 throughput of 19.66 TFLOPS is roughly 2.56 times higher than the B390's 7.680 TFLOPS. The texture rate of 614.4 GTexel/s is more than five times the B390's 120.0 GTexel/s. The pixel rate of 307.2 GPixel/s is more than five times the B390's 60.00 GPixel/s. The B770 has 32 ray tracing cores versus 12 on the B390, and its 256 TMUs compare to 48 on the B390.

A user who needs a self-contained graphics solution in a portable device should choose the B390, as it is the only one of the two that can operate without external power and with system-shared memory. A user who needs a discrete GPU with dedicated memory and high throughput should choose the B770, provided their system can support a dual-slot card with a 1x 6-pin and 1x 8-pin power requirement and a 550 W power supply. The B770 also offers display outputs (1x HDMI 2.1 and 3x DisplayPort 2.1), whereas the B390's display outputs are dependent on the portable device it is integrated into. The B770 is built on a 5 nm TSMC process, while the B390 uses Intel's 3 nm node, but this process difference does not translate into a performance advantage for the B390 in the recorded data.

Head-to-Head Benchmarks

There are no head-to-head benchmark entries between the Intel Arc B390 and the Intel Arc B770 in the database. The recorded wins count is 0 for each product, and the head-to-head array is empty. This absence of direct comparison data means that any quantitative comparison must rely on the individual benchmark scores and the specification fields.

The only benchmark score available is the B390's 3DMark Steel Nomad DX12 result of 1482. That score puts the B390 at the 9th percentile of all GPUs. The nearest rivals in the database are all NVIDIA parts, and the deltas are narrow. The B390 is 1.3% ahead of the GeForce GT 520MX, 1.5% ahead of the GeForce 800M, 2.5% ahead of the GeForce GT 625 OEM, and 2.7% ahead of the GeForce GT 710. These margins are small enough that they could be considered within run-to-run variance, but the consistent ordering suggests a real, if slight, performance edge over those specific parts.

The B770 has no benchmark scores at all. Its average benchmark score is recorded as 0, which is consistent with an empty benchmark list. Its percentile is listed as 50, which is the median, but this percentile appears to be a placeholder classification rather than a derived value from test results. Without a score, the database cannot confirm how the B770 performs relative to the B390 or any other GPU in actual workloads.

The specification comparison, however, provides a clear picture of the gap. The B770's FP32 compute of 19.66 TFLOPS is 2.56 times the B390's 7.680 TFLOPS. The FP16 compute of 39.32 TFLOPS is 2.56 times the B390's 15.36 TFLOPS. The texture fill rate of 614.4 GTexel/s is 5.12 times the B390's 120.0 GTexel/s. The pixel fill rate of 307.2 GPixel/s is 5.12 times the B390's 60.00 GPixel/s. The B770's memory bandwidth of 512.0 GB/s is not directly comparable to the B390 because the B390 uses system-shared memory with bandwidth that is system dependent. The B770's 16 GB of GDDR6 memory on a 256-bit bus contrasts with the B390's system-shared memory of unspecified size and type.

Clock speeds also differ. The B390 has a base clock of 300 MHz and a boost clock of 2500 MHz. The B770 has a base clock of 2100 MHz and a boost clock of 2400 MHz. The B390's boost clock is higher, but this does not compensate for the B770's much larger execution resource pool. The B770's base clock is seven times the B390's base clock, and its boost clock is 100 MHz lower than the B390's boost clock. The higher base clock on the B770 indicates that it is designed to sustain high throughput continuously, whereas the B390's low base clock and high boost clock suggest a power-constrained design that ramps up under load.

FAQ

Q: What is the performance difference between the Intel Arc B390 and the Intel Arc B770?

A: The B390 has a recorded 3DMark Steel Nomad DX12 score of 1482, while the B770 has no recorded benchmark scores. The specification data shows the B770 at 19.66 TFLOPS FP32 versus 7.680 TFLOPS for the B390, and 614.4 GTexel/s texture rate versus 120.0 GTexel/s.

Q: Which GPU has more memory?

A: The B770 has 16 GB of GDDR6 memory on a 256-bit bus with 512.0 GB/s bandwidth. The B390 uses system-shared memory, with the size, type, bus width, and bandwidth all listed as system dependent or system shared.

Q: How does the B390 compare to its nearest rivals?

A: The B390's score of 1482 puts it 1.3% ahead of the GeForce GT 520MX, 1.5% ahead of the GeForce 800M, 2.5% ahead of the GeForce GT 625 OEM, and 2.7% ahead of the GeForce GT 710.

Q: What are the power requirements for each GPU?

A: The B390 has a TDP of 80 W and requires no power connectors. The B770 has a TDP of 225 W and requires a 1x 6-pin plus 1x 8-pin power connector setup, with a suggested power supply of 550 W.

Q: What architectures do the two GPUs use?

A: The B390 uses the Xe3-LPG architecture on the Panther Lake chip, built on a 3 nm process at Intel. The B770 uses the Xe2-HPG architecture on the BMG-G31 chip, built on a 5 nm process at TSMC, with a die size of 368 mm².

Q: Are these GPUs compatible with the same systems?

A: No. The B390 is an IGP with an IGP bus interface, designed for portable devices. The B770 is a dual-slot discrete card with a PCIe 4.0 x16 interface and dedicated display outputs (1x HDMI 2.1 and 3x DisplayPort 2.1).

Architecture Differences

The two GPUs are built on different architectures, different process nodes, and different chips, which explains their divergent performance classes.

The Intel Arc B390 is based on the Panther Lake chip with the Xe3-LPG architecture. It belongs to the Arc Graphics-M generation for Panther Lake. The process node is 3 nm, and the foundry is Intel. The B390 uses an integrated design: the bus interface is IGP, the slot width is IGP, and the power connectors are none. Its memory subsystem is entirely system shared, meaning there is no dedicated VRAM. The memory size, type, bus width, and bandwidth are all listed as system shared or system dependent. The B390 has 1536 shading units, 48 texture mapping units, 24 raster operation units, and 12 ray tracing cores. Its base clock is 300 MHz and its boost clock is 2500 MHz. The pixel rate is 60.00 GPixel/s, the texture rate is 120.0 GTexel/s, and the FP32 compute is 7.680 TFLOPS. The FP16 compute is 15.36 TFLOPS with a 2:1 ratio. The TDP is 80 W. The display outputs are listed as portable device dependent, meaning the B390 does not have fixed display connectors of its own.

The Intel Arc B770 is based on the BMG-G31 chip with the Xe2-HPG architecture. It belongs to the Battlemage generation in the Arc 7 series. The process node is 5 nm, and the foundry is TSMC. The die size is 368 mm². The B770 is a discrete card with a PCIe 4.0 x16 bus interface and a dual-slot footprint. It uses 1x 6-pin plus 1x 8-pin power connectors and has a suggested power supply of 550 W. Its memory is 16 GB of GDDR6 on a 256-bit bus, with a bandwidth of 512.0 GB/s. The memory clock is 2000 MHz, 16 Gbps effective. The B770 has 4096 shading units, 256 TMUs, 128 ROPs, and 32 ray tracing cores. Its base clock is 2100 MHz and its boost clock is 2400 MHz. The pixel rate is 307.2 GPixel/s, the texture rate is 614.4 GTexel/s, FP32 compute is 19.66 TFLOPS, and FP16 compute is 39.32 TFLOPS with a 2:1 ratio. The TDP is 225 W. The display outputs are 1x HDMI 2.1 and 3x DisplayPort 2.1.

The B390's predecessor is not listed, and its successor is not listed. The B770's predecessor is listed as Alchemist, and its successor is not listed. The B390 has a production status of active, while the B770's production status is not specified in the database. The release date for the B390 is January 26, 2026, and the release date for the B770 is December 31, 2025. Neither product has a recorded launch MSRP.

Both GPUs support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. Neither has tensor cores listed in the database. The B390 has no transistor count or die size listed, and the B770's transistor count is also unknown, though its die size is 368 mm².

The architectural split is fundamental. The B390 uses a newer process node (3 nm versus 5 nm) and a newer architecture generation (Xe3-LPG versus Xe2-HPG), but it is designed for integration and low power. The B770 uses an older process node but a much larger die and a discrete design with dedicated memory. The B390's 12 ray tracing cores are less than half the B770's 32. The B390's 48 TMUs are less than a fifth of the B770's 256. The B390's 24 ROPs are less than a fifth of the B770's 128. These resource differences, combined with the B770's dedicated 512.0 GB/s memory bandwidth, create a performance gap that the B390's higher boost clock cannot close. The database records no head-to-head benchmarks, so the exact magnitude of the gap in real workloads remains unmeasured, but the specification data indicates the B770 is in a different performance tier entirely.