Intel Core Ultra 9 285HX vs Intel Core Ultra X9 378H Comparison

The Intel Core Ultra 9 285HX and the Intel Core Ultra X9 378H are two mobile processors from different generations of Intel’s Core Ultra lineup. The 285HX belongs to the Core Ultra Series 2, based on the Arrow Lake architecture, while the 378H is part of the Core Ultra Series 3, built on the Panther Lake platform. The recorded data shows a clear split: the 285HX dominates multi-threaded workloads, while the 378H delivers a significant single-core advantage in certain Cinebench tests. The 285HX records 15 benchmark wins out of 17 head-to-head comparisons, with the 378H taking only two.

FAQ

Q: Which processor has more cores and threads?

A: The Intel Core Ultra 9 285HX has 24 cores and 24 threads. The Intel Core Ultra X9 378H has 16 cores and 16 threads. Both processors lack hyper-threading, so thread counts match core counts.

Q: How do the two processors compare in average benchmark scores?

A: The 285HX has an average benchmark score of 76155, placing it in the 95th percentile of all CPUs. The 378H has an average score of 47468, putting it in the 89th percentile. The 285HX sits near processors like the AMD Ryzen 9 8945HX (76212, -0.1%) and the AMD Ryzen 9 9950X3D (75779, 0.5%). The 378H is comparable to the Intel Core i7-13700KF (47330, 0.3%) and the Intel Core i9-12900F (47176, 0.6%).

Q: What is the difference in boost clock speeds?

A: The 285HX has a boost clock of 5.50 GHz. The 378H has a boost clock of 5.00 GHz. The 285HX also has a higher base clock at 2.80 GHz versus 2.00 GHz for the 378H.

Q: Which processor supports ECC memory?

A: The Intel Core Ultra 9 285HX supports ECC memory. The Intel Core Ultra X9 378H does not support ECC memory.

Q: What are the memory bandwidth differences?

A: The 285HX supports DDR5 memory with a dual-channel bus and a recorded bandwidth of 102.4 GB/s. The 378H supports LPDDR5X memory with a dual-channel bus and a higher recorded bandwidth of 153.6 GB/s.

Q: When were both processors released?

A: The 285HX was released on 2025-01-12. The 378H was released on 2026-04-03. Both are listed as active in production status.

Architecture Differences

The two processors come from different Intel generations and use different silicon. The 285HX is built on the Arrow Lake architecture, specifically the Arrow Lake-HX codename, and uses a 3 nm process node from TSMC. The 378H uses the Panther Lake codename, also on a 3 nm process node, but the foundry is Intel. This distinction matters for manufacturing and design philosophy, though the data does not include transistor counts or die sizes for the 378H.

The 285HX has a significantly larger core configuration: 24 cores versus 16 cores for the 378H. Both have the same L1 cache size at 192 KB per core. The L2 cache differs, with the 285HX offering 3 MB per core compared to 2.5 MB per core on the 378H. The L3 cache is a major difference: the 285HX has 36 MB shared, while the 378H has only 18 MB shared. This doubling of L3 cache likely contributes to the 285HX’s multi-threaded performance advantage.

The integrated graphics differ substantially. The 285HX uses an Arc Xe-LPG Graphics solution with 64 execution units. The 378H uses an Arc B390 GPU. The data does not provide execution unit counts for the B390, but the architectural generation is different, which may affect graphics performance.

Memory support also separates the two. The 285HX uses DDR5, while the 378H uses LPDDR5X. The 378H has a higher memory bandwidth at 153.6 GB/s versus 102.4 GB/s for the 285HX. This suggests the 378H is designed for more power-efficient, integrated memory configurations, while the 285HX relies on a broader memory bus with a lower bandwidth figure.

PCIe connectivity is another divergence. The 285HX provides Gen 5 with 20 lanes (CPU only), while the 378H provides Gen 5 with only 4 lanes (CPU only). This makes the 285HX far more suited to systems requiring multiple high-speed expansion devices. The 378H’s limited lane count points to a more integrated, compact laptop design.

The 285HX has an unlocked multiplier, whereas the 378H does not. This gives the 285HX overclocking headroom for enthusiasts. The 285HX uses the Intel BGA 2114 socket, while the 378H uses the Intel BGA 2540 socket. These are physically different sockets, meaning they are not interchangeable in a motherboard.

The 285HX’s TDP is 55 watts, while the 378H’s TDP is 25 watts. This 30-watt difference reflects the 285HX’s higher core count and likely higher sustained power draw. The 378H is designed for lower power envelopes, which suits thinner laptops.

Where Each One Wins

The Intel Core Ultra 9 285HX wins in nearly every multi-threaded scenario. In Cinebench R15 multicore, it scores 5656.5 versus 3281 for the 378H, a delta of 72.4%. In Cinebench R20 multicore, the margin is 48% (20236 versus 13672). Even in Cinebench R23 multicore, where the gap narrows, the 285HX still wins by 11.9% with a score of 36429.5 versus 32553.

PassMark results reinforce this pattern. Data compression shows a 63.5% win for the 285HX (631885 versus 386591). Encryption is 62.8% higher (48567 versus 29840). Floating point math is 70.3% higher (194998 versus 114500). Integer math is 67.5% higher (155076 versus 92603). Multithreaded performance is 48.6% higher (56902 versus 38298). Random string sorting is 72.9% higher (77196 versus 44648). Find prime numbers is 28.9% higher (460 versus 357). Extended instructions are 56.9% higher (49148 versus 31315).

The 285HX also wins single-threaded tests in PassMark, although by smaller margins. PassMark single thread shows 4618 versus 4453, a 3.7% lead. The same margin appears in the PassMark singlethread test. Physics is a narrow 2.1% win (3476 versus 3404).

The Intel Core Ultra X9 378H wins only two tests, both in Cinebench single-core. In Cinebench R15 singlecore, the 378H scores 462 versus 323.5 for the 285HX, a 30% advantage. In Cinebench R23 singlecore, the 378H scores 4595 versus 2187.5, a 52.4% advantage. The 378H’s single-core strength in these two tests likely comes from its newer Panther Lake architecture and higher IPC per core, despite the lower clock speed.

The 378H also wins Cinebench R20 singlecore? No, the data shows the 285HX wins that test with 2856 versus 1929, a 48.1% delta. This inconsistency across single-core Cinebench versions suggests the 378H’s advantage is not universal. The 285HX leads in R20 singlecore, but the 378H leads in R15 and R23 singlecore. The 378H’s wins are real but confined to two specific test versions.

For workloads like video rendering, 3D modeling, or any task that scales with core count, the 285HX is the clear choice. For short single-threaded bursts in certain benchmarks, the 378H shows an edge, but that edge does not appear in PassMark single-thread or R20 singlecore.

Specification Differences

The core count difference is the most obvious spec split: 24 cores for the 285HX versus 16 for the 378H. Thread counts match cores for both, so no hyper-threading on either. Base and boost clocks differ, with the 285HX at 2.80 GHz base and 5.50 GHz boost, and the 378H at 2.00 GHz base and 5.00 GHz boost.

TDP is 55 watts for the 285HX and 25 watts for the 378H. Sockets differ: Intel BGA 2114 versus Intel BGA 2540. The 285HX uses TSMC as foundry, the 378H uses Intel. Both use a 3 nm process node.

Cache sizes differ in L2 and L3. L1 is the same at 192 KB per core. L2 is 3 MB per core for the 285HX versus 2.5 MB for the 378H. L3 is 36 MB shared versus 18 MB shared.

Memory support: DDR5 versus LPDDR5X. Both are dual-channel. Memory bandwidth: 102.4 GB/s versus 153.6 GB/s. ECC memory is supported only on the 285HX.

PCIe lanes: 20 lanes (Gen 5, CPU only) for the 285HX versus 4 lanes (Gen 5, CPU only) for the 378H. Integrated graphics: Arc Xe-LPG Graphics 64EU versus Arc B390. The 285HX has an unlocked multiplier; the 378H does not.

The 285HX has a part number of SRVFJ, while the 378H has an unknown part number. Release dates differ by over a year, with the 378H coming later. Both are active production parts.

Head-to-Head Benchmarks

The largest win for the 285HX comes in PassMark random string sorting, where it scores 77196 against 44648, a 72.9% delta. This is followed closely by Cinebench R15 multicore with a 72.4% delta (5656.5 versus 3281). Floating point math shows a 70.3% delta (194998 versus 114500). Integer math is 67.5% higher (155076 versus 92603). Data compression is 63.5% higher (631885 versus 386591). Data encryption is 62.8% higher (48567 versus 29840). Extended instructions are 56.9% higher (49148 versus 31315).

Cinebench R20 multicore shows a 48% delta (20236 versus 13672), and Cinebench R20 singlecore shows a 48.1% delta (2856 versus 1929). PassMark multithread is 48.6% higher (56902 versus 38298). Find prime numbers is 28.9% higher (460 versus 357). Cinebench R23 multicore is a smaller 11.9% delta (36429.5 versus 32553). PassMark single thread and singlethread both show a 3.7% delta (4618 versus 4453). PassMark physics is the narrowest 285HX win at 2.1% (3476 versus 3404).

The 378H’s two wins are both in Cinebench single-core tests. Cinebench R23 singlecore shows the largest margin for the 378H, with a 52.4% delta (4595 versus 2187.5). Cinebench R15 singlecore shows a 30% delta (462 versus 323.5). These two wins are substantial, but they do not offset the 285HX’s dominance across the other 15 tests.

The 285HX’s wins are not just numerous but also large in magnitude. Eight of its wins exceed 60% deltas, and several exceed 70%. The 378H’s wins are large as well, but they are isolated to single-core Cinebench workloads. In the broader PassMark suite, the 378H cannot match the 285HX even in single-threaded tests.

The data indicates that the 285HX is a high-core-count processor designed for demanding multi-threaded tasks, while the 378H is a lower-power part with a strong single-core showing in specific legacy benchmarks. For users prioritizing Cinebench R23 single-core performance, the 378H offers a notable advantage, but for almost any other measured workload, the 285HX delivers substantially higher scores. The 285HX’s 95th percentile ranking versus the 378H’s 89th percentile confirms its overall higher standing in the database.