Intel Core Ultra 5 235 vs Intel Core Ultra X7 358H Comparison

Head-to-Head Benchmarks

The benchmark data presents a clear split between the Intel Core Ultra 5 235 and the Intel Core Ultra X7 358H. The X7 358H dominates the Cinebench suite, while the Ultra 5 235 sweeps most PassMark workloads. The largest margin belongs to the X7 358H in Cinebench R15 multicore, where it scores 3027 versus 1488, a 50.8% advantage. That is a massive gap, one that suggests the mobile chip's core configuration and architecture handle heavily threaded rendering workloads far more effectively.

The Cinebench R20 multicore test shows a similar story. The X7 358H posts 12011 against 6202 for the Ultra 5 235, a 48.4% lead. The R20 single-core result follows the same pattern: 1695 versus 875, again 48.4% ahead. Even Cinebench R15 single-core favors the X7 358H, with 301.5 against 210, a 30.3% margin. The only Cinebench test where the Ultra 5 235 wins is R23 single-core, and it does so by a hair: 2085 versus 2080, a 0.2% delta. That is effectively a tie, but the database records it as a win for the desktop part.

The PassMark suite flips the script. The Ultra 5 235 wins 10 of the 12 PassMark tests recorded in the head-to-head table, with its largest edge in random string sorting at 48980 versus 40357, a 21.4% lead. Data compression favors the Ultra 5 235 by 17.5% (390711 versus 332508), and extended instructions show a 20.1% advantage (32752 versus 27274). Floating-point math goes to the Ultra 5 235 at 117951 versus 103842, a 13.6% margin, while integer math is closer at 5.8% (87948 versus 83147). Data encryption shows a 12.5% edge for the desktop chip (29293 versus 26046), and find prime numbers lands at 10.1% (371 versus 337). The multithread test gives the Ultra 5 235 an 11.9% win (37816 versus 33802), and single-thread tests show 4516 versus 4124, a 9.5% margin.

The X7 358H does take one PassMark test: physics, where it scores 3021 against 2570, a 14.9% win. That result is notable because the physics workload often correlates with memory latency and core scheduling, areas where the mobile chip's design may differ. Overall, the head-to-head table records 11 wins for the Ultra 5 235 and 6 for the X7 358H.

Where Each One Wins

The data indicates that the X7 358H is the stronger choice for Cinebench-style rendering workloads. Its multicore scores in R15, R20, and R23 are decisively higher, with the R23 multicore result at 18747 versus 14769, a 21.2% lead. Single-core Cinebench results also favor the mobile chip, except for the near-tie in R23. The physics test adds another win for the X7 358H, suggesting it handles certain simulation or rigid-body workloads better.

The Ultra 5 235, by contrast, wins across the broader PassMark suite. Compression, encryption, extended instructions, prime number finding, floating-point math, integer math, multithread, random string sorting, and single-thread tests all go to the desktop part. These workloads cover a wide range of everyday and professional tasks, from file compression to cryptographic operations to general arithmetic. The margins are moderate, mostly between 5.8% and 21.4%, but they are consistent. A user running mixed productivity tasks would likely see the Ultra 5 235 finish ahead in most of them.

The split is stark: rendering and physics favor the X7 358H, while nearly everything else in the PassMark suite favors the Ultra 5 235. The average benchmark scores reflect this divergence. The Ultra 5 235 posts an average benchmark score of 46062, while the X7 358H sits at 40967. The desktop chip also holds a higher percentile ranking among all CPUs at 89, compared to 87 for the mobile part.

Architecture Differences

The two processors come from different Intel lineups and use different foundries. The Ultra 5 235 is part of the Core Ultra Series 2, built on Arrow Lake with the codename Arrow Lake-S. It uses a 3 nm process from TSMC. The X7 358H belongs to the Core Ultra Series 3, codenamed Panther Lake, specifically Panther Lake-H. It also uses a 3 nm node, but the foundry is Intel rather than TSMC. That difference in manufacturing source may influence how each chip behaves under load, although the recorded data does not isolate this factor.

Core counts differ as well. The Ultra 5 235 has 14 cores and 14 threads, meaning no hyperthreading. The X7 358H has 16 cores and 16 threads, also without hyperthreading. The X7 358H therefore has two additional physical cores, which likely contributes to its multicore Cinebench dominance. Cache hierarchies are not identical. Both chips list 192 KB of L1 per core and 3 MB of L2 per core, but the shared L3 cache is larger on the Ultra 5 235 at 24 MB, compared to 18 MB on the X7 358H. That extra 6 MB of L3 may explain the desktop chip's advantage in data-heavy PassMark tests like compression and random string sorting.

Transistor counts and die sizes are only recorded for the Ultra 5 235: 17,800 million transistors on a 243 mm² die. The X7 358H has no such figures in the database. The integrated graphics differ too. The Ultra 5 235 uses Arc Xe-LPG Graphics with 24 execution units, while the X7 358H carries Arc B390. The memory support also diverges: the Ultra 5 235 supports DDR5, while the X7 358H uses LPDDR5X. Both are dual-channel, but the memory bandwidth is higher on the X7 358H at 153.6 GB/s, versus 102.4 GB/s for the Ultra 5 235.

Specification Differences

The socket and market segment set these chips apart fundamentally. The Ultra 5 235 is a desktop processor on Intel Socket 1851, while the X7 358H is a mobile chip on Intel BGA 2540. The TDP reflects that: the Ultra 5 235 draws 65 W, while the X7 358H is rated at 25 W. Despite the lower power envelope, the X7 358H manages higher multicore Cinebench scores, which points to a more efficient architecture or better memory bandwidth.

Clock speeds differ in an interesting way. The Ultra 5 235 has a base clock of 3.40 GHz and a boost clock of 5.00 GHz. The X7 358H has a much lower base clock of 1.90 GHz and a boost of 4.80 GHz. The higher base clock on the desktop chip does not translate into single-core Cinebench wins, except the marginal R23 result. PCIe lane counts also vary significantly. The Ultra 5 235 provides Gen 5 with 20 lanes from the CPU, while the X7 358H offers Gen 5 with only 4 lanes. This makes the desktop part far more suitable for multi-GPU or high-expansion builds, though the database does not include any expansion tests.

The release dates are recorded as 2025-01-06 for the Ultra 5 235 and 2026-01-04 for the X7 358H. The Ultra 5 235 has a launch MSRP of $257, while the X7 358H has no recorded launch MSRP. Both parts are active in production, and neither has an unlocked multiplier. The part numbers are SRQAS for the Ultra 5 235 and SA4RAQ9ET for the X7 358H. Neither supports ECC memory.

FAQ

Q: Which processor wins more head-to-head benchmarks?

A: The Intel Core Ultra 5 235 wins 11 of the 17 recorded head-to-head tests, while the Intel Core Ultra X7 358H wins 6.

Q: What is the largest benchmark margin between the two?

A: The largest margin is in Cinebench R15 multicore, where the Intel Core Ultra X7 358H scores 3027 against 1488, a 50.8% lead.

Q: Does the Intel Core Ultra 5 235 win any Cinebench test?

A: Yes, it wins Cinebench R23 single-core with a score of 2085 versus 2080, a 0.2% margin.

Q: How do their average benchmark scores compare?

A: The Intel Core Ultra 5 235 has an average benchmark score of 46062, while the Intel Core Ultra X7 358H has an average of 40967.

Q: What memory bandwidth does each chip support?

A: The Intel Core Ultra 5 235 supports 102.4 GB/s with DDR5, while the Intel Core Ultra X7 358H supports 153.6 GB/s with LPDDR5X.

Q: Which chip has more cores?

A: The Intel Core Ultra X7 358H has 16 cores and 16 threads, while the Intel Core Ultra 5 235 has 14 cores and 14 threads.

The Verdict

The data supports a clear split based on workload type. For Cinebench rendering and physics simulation, the Intel Core Ultra X7 358H is the stronger part. Its multicore R15 score is more than double the Ultra 5 235, and its R20 multicore lead is nearly double. Even in single-core Cinebench tests, except the R23 near-tie, the X7 358H comes out ahead. The mobile chip achieves this with a 25 W TDP, which is remarkable given the desktop chip runs at 65 W.

For general productivity and the PassMark suite, the Intel Core Ultra 5 235 is the better performer. It wins compression, encryption, extended instructions, prime numbers, floating-point math, integer math, multithread, random string sorting, and single-thread tests. Its margins are not enormous, but they are consistent across a broad range of workloads. The higher average benchmark score of 46062 versus 40967 confirms that the desktop chip offers more overall compute throughput in the database's aggregate metrics.

The choice comes down to the intended environment. A desktop build with expansion needs calls for the Ultra 5 235, which provides 20 PCIe Gen 5 lanes, a 24 MB L3 cache, and a 5.00 GHz boost clock. A mobile system or a compact chassis with modest power delivery favors the X7 358H, which delivers superior Cinebench results and higher memory bandwidth at 153.6 GB/s. The X7 358H also uses Intel as the foundry, whereas the Ultra 5 235 uses TSMC. Both chips share the same 3 nm process node, but the architectural lineage differs: Arrow Lake-S versus Panther Lake-H. The recorded data does not declare an overall winner, but it does define distinct territories. Rendering leans toward the X7 358H, and everything else leans toward the Ultra 5 235.