Intel Core Ultra 7 265HX vs Intel Core Ultra X7 358H Comparison

Head-to-Head Benchmarks

The Intel Core Ultra 7 265HX dominates the head-to-head comparison, winning 16 of the 17 recorded benchmarks. The single exception is PassMark physics, where the Core Ultra X7 358H takes a narrow 1.4% lead. Every other test shows substantial margins in favor of the 265HX, often exceeding 50%.

The most decisive victory comes in Cinebench R23 single-core, where the 265HX scores 5737 against 2080, a 175.8% advantage. This is the largest delta recorded between the two processors. Multi-core performance in the same test shows a 116.8% lead, with the 265HX posting 40642 versus 18747. These results indicate a fundamental throughput gap that extends across both lightly threaded and heavily threaded workloads.

Cinebench R15 and R20 tell a similar story. In R15 multi-core, the 265HX records 4096 against 3027, a 35.3% edge. Single-core in R15 shows a 91.7% lead, 578 versus 301.5. R20 multi-core delivers a 42.1% advantage, 17069 against 12011, and single-core matches that 42.1% delta with 2409 versus 1695.

PassMark results reinforce the pattern. Data compression favors the 265HX by 53.9%, 511817 against 332508. Encryption shows a 51.5% lead, 39472 versus 26046. Extended instructions come in at 49.4% ahead, 40741 against 27274. Integer math delivers a 52.7% margin, 126954 versus 83147, while floating-point math shows a 55.6% edge, 161605 against 103842. Random string sorting lands at 54.8% ahead, 62458 versus 40357. Prime number finding is the closest PassMark result, with the 265HX ahead by 20.5%, 406 versus 337.

Multi-threaded PassMark performance shows a 42% advantage, 47985 versus 33802. Single-thread PassMark is closer, with the 265HX leading by 9.1%, 4500 versus 4124. The physics test is the only reversal, with the 358H posting 3021 against 2978, a 1.4% margin that falls within noise but still counts as a win.

The average benchmark score reflects this overall disparity. The 265HX averages 63173, placing it in the 93rd percentile of all CPUs in the database. The 358H averages 40967, which lands in the 87th percentile. The nearest rivals corroborate the positioning. The 265HX sits within 0.7% of the Intel Core Ultra 7 255HX, 0.5% of the AMD Ryzen AI Embedded P185, 0.2% below the AMD Ryzen AI 7 450G, and 0.1% above the Intel Core i7-13790F. The 358H sits within 0.7% of the Intel Core Ultra 7 366H, 0.6% of the AMD Ryzen AI 5 PRO 440, 0.6% of the Intel Core Ultra 7 356H, and 0.6% above the AMD Ryzen AI 5 PRO 435G.

The Verdict

The data points to one clear conclusion: the Intel Core Ultra 7 265HX is the superior processor for nearly every workload category measured. Its 16 wins versus 1, combined with an average score 54.2% higher, establish a decisive performance hierarchy.

The 265HX delivers massive multi-core advantages, exceeding 100% in Cinebench R23 multi-core and maintaining 35% to 55% leads across all other multi-threaded tests. Content creation, compilation, data processing, and heavy multitasking all favor this chip by wide margins. The single-core advantage is equally pronounced, with Cinebench R15 and R23 showing 91.7% and 175.8% leads respectively, while PassMark single-thread shows a more modest 9.1% edge.

The 358H offers only one measurable advantage: a 1.4% lead in PassMark physics. This narrow win does not offset the substantial deficits elsewhere. The 358H also consumes less power according to the TDP figures, but the database does not record any performance-per-watt benchmarks, so no efficiency conclusion can be drawn from the measured data.

For buyers prioritizing raw benchmark performance, the 265HX is the clear selection. The 358H remains a functional mobile processor, but every recorded benchmark except one favors its competitor.

Architecture Differences

The two processors diverge significantly in architecture despite sharing the same 3 nm process node. The 265HX uses Arrow Lake architecture with the Arrow Lake-HX codename, fabricated by TSMC. The 358H uses Panther Lake architecture with the Panther Lake-H codename, fabricated by Intel's own foundry. The 265HX belongs to Core Ultra Series 2, while the 358H belongs to Core Ultra Series 3.

Core configuration differs substantially. The 265HX packs 20 cores and 20 threads, while the 358H has 16 cores and 16 threads. Neither processor uses simultaneous multithreading, so core and thread counts are identical for each. The 265HX also carries an unlocked multiplier, whereas the 358H is locked.

Cache hierarchies share the same per-core L1 and L2 allocations, 192 KB and 3 MB respectively. The shared L3 cache differs: the 265HX has 30 MB, while the 358H has 18 MB. This 12 MB L3 difference contributes to the performance gap in cache-sensitive workloads.

Memory support separates the two as well. The 265HX supports DDR5 over a dual-channel bus with 102.4 GB/s bandwidth. The 358H supports LPDDR5X over a dual-channel bus with 153.6 GB/s bandwidth. The 358H therefore has a 50% higher theoretical memory bandwidth, yet this does not translate into benchmark wins. PCIe lane allocation also differs: the 265HX provides Gen 5 with 20 CPU lanes, while the 358H provides Gen 5 with 4 CPU lanes.

Integrated graphics differ. The 265HX uses Arc Xe-LPG Graphics with 64 execution units. The 358H uses Arc B390. The database does not record integrated graphics benchmark scores, so no performance comparison is possible from the measured data.

Clock speeds favor the 265HX. It has a base clock of 2.60 GHz and a boost clock of 5.30 GHz. The 358H has a base clock of 1.90 GHz and a boost clock of 4.80 GHz. The TDP also differs, with the 265HX rated at 55 and the 358H at 25. Sockets are incompatible: the 265HX uses Intel BGA 2114, while the 358H uses Intel BGA 2540.

Release dates show the 358H as a later product, with the 265HX released in January 2025 and the 358H in January 2026. Both remain in active production.

FAQ

Q: Which processor is faster in multi-core workloads?

A: The Intel Core Ultra 7 265HX wins every multi-core benchmark. Cinebench R23 multi-core shows a 116.8% lead, R20 shows 42.1%, R15 shows 35.3%, and PassMark multi-thread shows 42%. The 265HX also leads in data compression by 53.9%, encryption by 51.5%, integer math by 52.7%, and floating-point math by 55.6%.

Q: Does the Intel Core Ultra X7 358H win any benchmark?

A: Yes, one. The 358H wins PassMark physics with a score of 3021 against 2978, a 1.4% margin. This is the only recorded victory for the 358H across all 17 head-to-head tests.

Q: How do the single-core scores compare?

A: The 265HX leads in every single-core test. Cinebench R23 single-core shows a 175.8% advantage, R15 shows 91.7%, R20 shows 42.1%, and PassMark single-thread shows 9.1%. The 265HX scores 5737 in R23 and 4500 in PassMark, while the 358H scores 2080 and 4124 respectively.

Q: What is the difference in core and thread counts?

A: The 265HX has 20 cores and 20 threads. The 358H has 16 cores and 16 threads. Neither processor supports simultaneous multithreading, so each core corresponds to exactly one thread.

Q: Which processor has more L3 cache?

A: The 265HX has 30 MB of shared L3 cache. The 358H has 18 MB. Both allocate 192 KB of L1 cache and 3 MB of L2 cache per core.

Q: Do the processors support the same memory types?

A: No. The 265HX supports DDR5 with 102.4 GB/s bandwidth. The 358H supports LPDDR5X with 153.6 GB/s bandwidth. Both use dual-channel memory buses, and neither supports ECC memory.

Where Each One Wins

The Intel Core Ultra 7 265HX wins across the entire benchmark suite except for one test. Its strengths are most pronounced in Cinebench workloads, where multi-core and single-core scores both exceed 100% leads in R23. The 35.3% R15 multi-core margin and 42.1% R20 margins indicate consistent advantages across rendering and simulation tasks. PassMark integer and floating-point math, with 52.7% and 55.6% leads, point to strong general computation performance. Data compression and encryption, ahead by 53.9% and 51.5%, suggest the 265HX handles file archiving, database workloads, and security operations with significant headroom.

The single PassMark physics win for the 358H, at 1.4%, is too narrow to indicate a meaningful advantage in physics simulation. The 358H also shows a closer single-thread PassMark result, trailing by only 9.1%, which narrows the gap in lightly threaded general-purpose tasks. However, the Cinebench R15 single-core deficit of 91.7% and R23 single-core deficit of 175.8% show that the 358H cannot match the 265HX in single-threaded rendering performance.

For workloads dominated by the recorded benchmarks, the 265HX is the pick. The 358H does not offer a compelling use case based on the measured data, since its only win is a sub-2% physics margin. Users who prioritize the specific physics test may find the 358H marginally better, but the broader benchmark picture heavily favors the 265HX.

Specification Differences

| Specification | Intel Core Ultra 7 265HX | Intel Core Ultra X7 358H |

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

| Series | Core Ultra Series 2 | Core Ultra Series 3 |

| Cores | 20 | 16 |

| Threads | 20 | 16 |

| Base clock | 2.60 GHz | 1.90 GHz |

| Boost clock | 5.30 GHz | 4.80 GHz |

| TDP | 55 | 25 |

| Socket | Intel BGA 2114 | Intel BGA 2540 |

| Architecture | Arrow Lake | Not recorded |

| Codename | Arrow Lake-HX | Panther Lake |

| Process node | 3 nm | 3 nm |

| Foundry | TSMC | Intel |

| L3 cache | 30 MB (shared) | 18 MB (shared) |

| Memory support | DDR5 | LPDDR5X |

| Memory bandwidth | 102.4 GB/s | 153.6 GB/s |

| PCIe | Gen 5, 20 Lanes (CPU only) | Gen 5, 4 Lanes (CPU only) |

| Integrated graphics | Arc Xe-LPG Graphics 64EU | Arc B390 |

| Multiplier unlocked | Yes | No |

| Release date | 2025-01-12 | 2026-01-04 |

| Part number | SRVFH | SA4RAQ9ET |

The 265HX holds advantages in core count, clock speeds, L3 cache, PCIe lane count, and unlocked multiplier. The 358H counters with higher memory bandwidth, a different memory type, and a lower TDP. The process node is identical at 3 nm, but the foundries differ, with TSMC producing the 265HX and Intel producing the 358H. Both processors use 192 KB L1 and 3 MB L2 per core, neither supports ECC memory, and both target the mobile market segment.