Intel Core Ultra 9 285K vs Intel Core Ultra 9 290HX Plus Comparison

Head-to-Head Benchmarks

The head-to-head data shows a clean sweep: the Intel Core Ultra 9 285K wins all 17 recorded comparisons against the Intel Core Ultra 9 290HX Plus. No test goes to the mobile part. The margins, however, vary meaningfully by workload, and that spread tells the real story.

The largest single victory comes in PassMark extended instructions, where the 285K scores 62,277 against 51,290 for the 290HX Plus, a 21.4% advantage. This suggests the desktop chip sustains higher throughput on AVX-style workloads, a pattern that carries over to data compression. There, the 285K posts 790,052 versus 658,724, a 19.9% lead. Random string sorting follows closely at 18.2% ahead (94,927 versus 80,327), and physics simulation shows a 16.3% gap (3,938 versus 3,387). These are the workloads where the 285K's higher sustained power envelope appears to pay off most.

Data encryption lands at 15.5% ahead (57,745 versus 50,008). Cinebench R20 multicore and PassMark multithread both show 13.2% gaps, with the 285K scoring 24,003 versus 21,198 and 67,260 versus 59,439 respectively. The Cinebench R20 single-core result also hits 13.2% (3,388 versus 2,992), which is notable because single-core gaps are typically smaller. Floating-point math sits at 11.2% (224,324 versus 201,773).

Smaller but still consistent edges appear in Cinebench R15 multicore (8.6%, 6,494 versus 5,981), Cinebench R23 multicore (7.2%, 42,522 versus 39,684), and Cinebench R15 single-core (5.6%, 359 versus 340). The narrowest wins are in PassMark integer math (4.6%, 172,379 versus 164,839), PassMark find prime numbers (4.2%, 541 versus 519), PassMark single-thread (2.7%, 5,087 versus 4,951), and Cinebench R23 single-core (0.9%, 2,377 versus 2,356).

The single-thread results deserve extra attention. A 0.9% gap in Cinebench R23 single-core is nearly negligible, and the 2.7% PassMark single-thread gap is small. This means the 290HX Plus is not far behind in lightly threaded tasks, even though it trails badly in multithreaded and vector-heavy workloads. The desktop part’s advantage grows precisely when the workload scales across cores or stresses the memory system.

Architecture Differences

Both processors share the same fundamental silicon. Each has 24 cores and 24 threads, with no hyperthreading. Both use a 3 nm process node from TSMC, pack 17,800 million transistors on a 243 mm² die, and feature identical cache hierarchies: 192 KB of L1 per core, 3 MB of L2 per core, and 36 MB of shared L3. Memory support is DDR5 with a dual-channel bus and 102.4 GB/s bandwidth. ECC memory is enabled on both, and both carry the Arc Xe-LPG Graphics 64EU integrated GPU plus Gen 5 PCIe with 20 CPU lanes. The multiplier is unlocked on both parts, and both are listed as Active in production.

The differences start with the socket. The 285K uses Intel Socket 1851, a desktop platform, while the 290HX Plus uses Intel BGA 2114, a soldered mobile package. That alone explains much of the performance gap, because the desktop socket allows a 125 W TDP versus the mobile part’s 55 W TDP. The clock speeds reflect that thermal budget. The 285K has a 3.70 GHz base clock and a 5.70 GHz boost clock. The 290HX Plus runs at 2.70 GHz base and 5.50 GHz boost. The desktop chip’s 1 GHz base-clock advantage is substantial, and its 200 MHz boost advantage is smaller but still present.

The architecture codenames differ as well. The 285K is Arrow Lake-S, while the 290HX Plus is an Arrow Lake-HX Refresh. The 285K’s generation is listed as Ultra 9 (Arrow Lake), and the 290HX Plus as Ultra 9 (Arrow Lake-HX). The desktop part launched on 2024-10-23 with a launch MSRP of $589. The mobile part has a later release date of 2026-03-16 and no launch MSRP recorded. The part numbers also differ: SRQD5 for the 285K, SADSS for the 290HX Plus.

One notable symmetry: both parts have the same transistor count, die size, and cache layout. The performance difference is not a matter of more silicon on the desktop chip. It is a matter of how that silicon is powered and cooled. The 285K can draw more than twice the TDP, and the data shows that extra headroom translates directly into higher sustained scores.

FAQ

Q: Which CPU wins more benchmarks?

A: The Intel Core Ultra 9 285K wins all 17 recorded head-to-head tests. The Intel Core Ultra 9 290HX Plus does not win a single comparison.

Q: How large is the performance gap in multithreaded workloads?

A: The gap ranges from 7.2% in Cinebench R23 multicore to 13.2% in Cinebench R20 multicore and PassMark multithread. The largest multithreaded margin is 19.9% in PassMark data compression.

Q: Is the 290HX Plus competitive in single-core tasks?

A: It is close but still behind. The 285K leads by 5.6% in Cinebench R15 single-core, 13.2% in Cinebench R20 single-core, 0.9% in Cinebench R23 single-core, and 2.7% in PassMark single-thread.

Q: Do both CPUs have the same core count and cache?

A: Yes. Both have 24 cores, 24 threads, 192 KB L1 per core, 3 MB L2 per core, and 36 MB shared L3.

Q: Why does the 285K perform better despite identical silicon?

A: The 285K has a 125 W TDP, a 3.70 GHz base clock, and a 5.70 GHz boost clock. The 290HX Plus has a 55 W TDP, a 2.70 GHz base clock, and a 5.50 GHz boost clock. The higher power and clock rates allow the desktop part to sustain higher throughput.

Q: What is the market positioning of each chip?

A: The 285K is a desktop processor on Intel Socket 1851, released 2024-10-23. The 290HX Plus is a mobile processor on Intel BGA 2114, released 2026-03-16.

Specification Differences

| Field | Intel Core Ultra 9 285K | Intel Core Ultra 9 290HX Plus |

|-------|--------------------------|-------------------------------|

| Base clock | 3.70 GHz | 2.70 GHz |

| Boost clock | 5.70 GHz | 5.50 GHz |

| TDP | 125 W | 55 W |

| Socket | Intel Socket 1851 | Intel BGA 2114 |

| Codename | Arrow Lake-S | Arrow Lake-HX Refresh |

| Generation | Ultra 9 (Arrow Lake) | Ultra 9 (Arrow Lake-HX) |

| Market segment | Desktop | Mobile |

| Release date | 2024-10-23 | 2026-03-16 |

| Launch MSRP | $589 | None recorded |

| Part number | SRQD5 | SADSS |

All other recorded specifications match: 24 cores, 24 threads, 3 nm process, TSMC foundry, 17,800 million transistors, 243 mm² die, identical cache sizes, DDR5 memory, dual-channel bus, 102.4 GB/s bandwidth, ECC support, Gen 5 PCIe with 20 lanes, Arc Xe-LPG Graphics 64EU, and an unlocked multiplier.

The Verdict

The data points to a straightforward conclusion: the Intel Core Ultra 9 285K is the faster processor across every measured benchmark. Its 17 to 0 win record is unambiguous. The largest advantages appear in extended instructions (21.4%), data compression (19.9%), random string sorting (18.2%), and physics (16.3%). These are not marginal wins; they represent a meaningful performance tier separation.

The 290HX Plus is not a weak chip in absolute terms. It holds a 95th percentile ranking among all CPUs, and its average benchmark score of 79,574 places it just 0.3% behind the Intel Core i9-14900KF. The 285K, by comparison, sits at the 96th percentile with an average score of 83,807. The desktop part also edges out the Intel Core Ultra 9 290K Plus by 0.2% in average score, while the mobile chip trails that same rival by a larger margin.

The real differentiator is power. The 285K’s 125 W TDP versus the 290HX Plus’s 55 W TDP is the single largest architectural gap. The 1 GHz base-clock difference compounds that effect. For anyone building a desktop system with adequate cooling, the 285K is the clear choice. For anyone constrained to a mobile platform, the 290HX Plus is the only option of the two, and its 95th percentile ranking shows it is still a high-end part in that context.

Where Each One Wins

The Intel Core Ultra 9 285K wins in every recorded category, but the size of the win varies by workload type. The biggest margins cluster around SIMD-heavy and data-processing tasks. Extended instructions at 21.4% ahead, data compression at 19.9%, and random string sorting at 18.2% indicate that the desktop chip excels when the workload is vectorized or memory-bandwidth sensitive. Physics simulation at 16.3% and data encryption at 15.5% reinforce this pattern. These are the workloads where the 285K’s higher sustained clocks and power headroom matter most.

The 285K also wins decisively in mixed multithreaded benchmarks. Cinebench R20 multicore and PassMark multithread both show 13.2% leads, and Cinebench R15 multicore sits at 8.6%. The Cinebench R23 multicore gap narrows to 7.2%, but it is still a clear win. Integer math and prime-number finding show smaller edges of 4.6% and 4.2%, respectively, suggesting that scalar integer workloads are less sensitive to the power difference.

The 290HX Plus’s best showing is in Cinebench R23 single-core, where it trails by only 0.9%. PassMark single-thread is also close at 2.7%. In lightly threaded tasks that do not stress the power envelope, the mobile chip nearly keeps pace. The 290HX Plus also has the advantage of being a mobile part with a 55 W TDP, which makes it available in laptops where the 285K cannot be installed at all. For users who need a high-core-count CPU in a portable form factor, the 290HX Plus is the only one of these two that fits. But for raw performance in any workload where the CPU can stretch its legs, the 285K is the winner without exception.