Intel Core 5 223PE vs Intel Core Ultra 7 265H Comparison

The Intel Core Ultra 7 265H and the Intel Core 5 223PE are two processors that land remarkably close together in the database's aggregate rankings, yet they arrive from entirely different directions. One is a mobile part built on TSMC's 3 nm process; the other is a desktop chip etched on Intel's own 10 nm node. Their average benchmark scores sit within roughly two and a half percent of each other, and their percentile placements against all recorded CPUs differ by a single point, 88 versus 87. That proximity invites a closer question: when two chips this different score this similarly, what does each one actually trade away to get there? The head-to-head record answers it, with the Core Ultra 7 265H winning 12 of 17 measured tests and the Core 5 223PE taking 5, but the distribution of those wins tells a more interesting story than the tally alone.

Where Each One Wins

The Core Ultra 7 265H dominates in throughput-oriented and numerical workloads. It wins multicore rendering in Cinebench R15 by 12.1 percent and R20 by 9.2 percent, and its Passmark multithread score of 34027 beats the 223PE's 31124 by 9.3 percent. The gaps widen dramatically in pure computation: floating point math favors the 265H by 42.7 percent, data encryption by 41 percent, and prime number finding by 110.7 percent, the single largest margin in the entire dataset. String sorting, extended instructions, and Passmark single-thread also fall to the mobile chip.

The Core 5 223PE's wins cluster in a puzzling pattern. It takes Cinebench R23 single-core by 44.3 percent and R23 multicore by 24.6 percent, plus Cinebench R15 single-core by 18.4 percent, integer math by 14.4 percent, and data compression by 3.4 percent. That R23 sweep is striking given that the 265H wins the older R15 and R20 multicore tests outright. A result that flips direction between benchmark generations is worth questioning: the recorded data shows the 223PE at 3734 in R23 single-core versus 1568 in R20 single-core, a spread that suggests its Bartlett Lake architecture benefits disproportionately from how R23 exercises the cores. Whatever the cause, the desktop chip owns modern Cinebench, while the 265H owns Passmark's single-thread metric by 2.7 percent. Physics is effectively a draw, separated by 0.2 percent.

Architecture Differences

These two chips could hardly be less alike under the hood. The Core Ultra 7 265H is an Arrow Lake-H mobile processor manufactured by TSMC on a 3 nm node, part of Core Ultra Series 2, and it brings 16 cores and 16 threads to the table. The Core 5 223PE is a Bartlett Lake desktop processor built by Intel on a 10 nm process, with 8 cores and 16 threads. Notice the thread count: both chips offer 16 threads, but the 265H does it without hyperthreading-style doubling, which helps explain why its multicore wins come despite identical thread capacity.

Cache hierarchies diverge too. The 265H carries 192 KB of L1 per core and 3 MB of L2 per core; the 223PE carries 80 KB of L1 per core and 2 MB of L2 per core. Both share an identical 24 MB of L3, so the mobile chip's advantage lives in the faster tiers closest to each core. Memory support also splits them: the 265H accepts DDR5 and LPDDR5X at 102.4 GB/s of bandwidth over a dual-channel bus, while the 223PE accepts both DDR4 and DDR5 at 89.6 GB/s. The flexibility of DDR4 support is unique to the desktop part, but the bandwidth crown belongs to the mobile chip.

Platform characteristics differ sharply. The 265H is a BGA 2049 soldered mobile part with a 28 W TDP, while the 223PE sits in Socket 1700 at 65 W. The 223PE offers 16 CPU PCIe 5.0 lanes versus the 265H's 8, a meaningful difference for expansion. Integrated graphics also separate them: Arc Graphics 140T on the 265H against UHD Graphics 730 on the 223PE. Both support ECC memory, both have locked multipliers, and both are active products. The 223PE carries a launch MSRP of $232; no comparable figure is recorded for the 265H.

Head-to-Head Benchmarks

Walking through the numbers, the rendering story inverts across Cinebench generations. In R15 multicore, the 265H scores 2989 against 2666, a 12.1 percent win. In R20 multicore, it is 12131 against 11111, 9.2 percent. Then R23 flips: the 223PE posts 26455 against 19940, winning by 24.6 percent. Single-core follows the same shape, with the 265H ahead in R20 single-core by 9.2 percent (1712 versus 1568), but the 223PE ahead in R15 single-core by 18.4 percent and in R23 single-core by a massive 44.3 percent. The 223PE's R23 single-core score of 3734 is by far its strongest headline number.

Passmark tells the opposite story. Single-thread goes to the 265H, 4334 to 4219, a 2.7 percent edge. Multithread goes to the 265H as well, 34027 to 31124. The computational subtests then break open: floating point math at 109123 versus 76468 is a 42.7 percent gap, data encryption at 26005 versus 18448 is 41 percent, and find prime numbers at 335 versus 159 is a 110.7 percent blowout. Integer math is the 223PE's counterpunch at 99819 versus 85479, a 14.4 percent win, and data compression favors it by 3.4 percent (346623 versus 334711). Random string sorting favors the 265H by 13.8 percent, extended instructions by 8.6 percent, and physics is a near-tie at 2497 versus 2493.

Context from the rivals list sharpens the picture. The 265H's average score of 41621 places it 0.1 percent below the Intel Core 7 251TE, 0.1 percent above the Core i7-14650HX, 0.4 percent below the Core i7-12850HX, and 0.6 percent above the AMD Ryzen 9 5900X, meaning it competes in a dense cluster of established performance chips. The 223PE's average of 40585 puts it within a tenth of a percent of both the Intel Core 7 253PE and the Xeon 6357P, 0.2 percent ahead of the Core Ultra X7 368H, and 0.3 percent behind the AMD Ryzen AI 5 PRO 435G. Both chips sit in crowded territory.

The Verdict

The data points to a clear division of labor. For sustained parallel workloads, floating point computation, encryption, and anything that rewards many physical cores and deep per-core cache, the Core Ultra 7 265H is the stronger choice, and it does all of this within a 28 W mobile envelope on a soldered BGA package. For workloads that lean on single-threaded rendering performance in the newest Cinebench generation, integer math, or DDR4 compatibility in a socketed desktop with 16 PCIe 5.0 lanes, the Core 5 223PE is the pick. The aggregate scores are nearly identical, so the decision should rest on workload profile rather than the average: a 42.7 percent floating point gap or a 44.3 percent single-core rendering gap will overwhelm any two percent difference in blended averages.

FAQ

Q: Which chip is faster overall?

A: The Core Ultra 7 265H wins 12 of 17 head-to-head tests and posts the higher average benchmark score, 41621 versus 40585, but the Core 5 223PE wins several tests decisively, including Cinebench R23 single-core by 44.3 percent.

Q: How do their core counts compare?

A: The 265H has 16 cores and 16 threads; the 223PE has 8 cores and 16 threads. Both therefore expose 16 threads to software.

Q: Which processor has the better single-core performance?

A: It depends on the test. The 265H leads Passmark single-thread by 2.7 percent and Cinebench R20 single-core by 9.2 percent, while the 223PE leads Cinebench R23 single-core by 44.3 percent and R15 single-core by 18.4 percent.

Q: Do both support ECC memory?

A: Yes, both the Core Ultra 7 265H and the Core 5 223PE support ECC memory.

Q: What memory does each support?

A: The 265H supports DDR5 and LPDDR5X with 102.4 GB/s of bandwidth. The 223PE supports DDR4 and DDR5 with 89.6 GB/s of bandwidth.

Q: Which has the bigger biggest margin win in the dataset?

A: The 265H's find prime numbers score of 335 versus 159 is a 110.7 percent margin, the largest gap recorded between the two.

Specification Differences

| Field | Intel Core Ultra 7 265H | Intel Core 5 223PE |

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

| Market segment | Mobile | Desktop |

| Architecture | Arrow Lake | null (Bartlett Lake codename) |

| Process node | 3 nm | 10 nm |

| Foundry | TSMC | Intel |

| Cores / Threads | 16 / 16 | 8 / 16 |

| Base clock | 2.20 GHz | 2.90 GHz |

| Boost clock | 5.30 GHz | 5.20 GHz |

| TDP | 28 W | 65 W |

| Socket | Intel BGA 2049 | Intel Socket 1700 |

| L1 cache (per core) | 192 KB | 80 KB |

| L2 cache (per core) | 3 MB | 2 MB |

| Memory support | DDR5, LPDDR5X | DDR4, DDR5 |

| Memory bandwidth | 102.4 GB/s | 89.6 GB/s |

| PCIe | Gen 5, 8 lanes (CPU only) | Gen 5, 16 lanes (CPU only) |

| Integrated graphics | Arc Graphics 140T | UHD Graphics 730 |

| Release date | 2025-01-12 | 2026-03-08 |

| Launch MSRP | not recorded | $232 |

| Part number | SRQAQ | SA4QF |

Both chips share 24 MB of shared L3 cache, dual-channel memory buses, ECC support, locked multipliers, and active production status.