Intel Core Ultra 5 235H vs Intel Core Ultra 5 235T Comparison

Intel
INTEL

Intel Core Ultra 5 235H

CORE STATE Arrow Lake-H
CORE SPECS 14 Cores / 14 Threads
CLOCK SPEED 2.4 Base / 5 GHz Turbo
CACHE 18 MB (shared)
MAX TDP 28W
ARCHITECTURE Arrow Lake
nm
PROCESS 3 nm
LAUNCH DATE 2025
VS
Intel
INTEL

Core Ultra 5 235T

CORE STATE Arrow Lake-S
CORE SPECS 14 Cores / 14 Threads
CLOCK SPEED 2.2 Base / 5 GHz Turbo
CACHE 24 MB (shared)
MAX TDP 65W
ARCHITECTURE Arrow Lake
nm
PROCESS 3 nm
LAUNCH DATE 2025

PERFORMANCE BENCHMARKS

cinebench_cinebench_r15_multicore
2,580
2,644
cinebench_cinebench_r15_singlecore
364
373
cinebench_cinebench_r20_multicore
10,751
11,017
cinebench_cinebench_r20_singlecore
1,517
1,555
cinebench_cinebench_r23_multicore
25,598
26,232
cinebench_cinebench_r23_singlecore
3,613
3,703
passmark_data_compression
301,979
295,100
passmark_data_encryption
23,121
23,457
passmark_extended_instructions
23,354
23,212
passmark_find_prime_numbers
239
318
passmark_floating_point_math
93,509
106,546
passmark_integer_math
74,247
84,244
passmark_multithread
30,091
30,918
passmark_physics
1,985
2,157
passmark_random_string_sorting
36,208
35,377
passmark_single_thread
4,359
4,339
passmark_singlethread
4,359
4,339

Analysis: Intel Core Ultra 5 235H vs Intel Core Ultra 5 235T

Where Each One Wins

The benchmark data splits these two Arrow Lake parts into clear territories. The Intel Core Ultra 5 235T wins 12 of the 17 recorded head-to-head comparisons, while the Intel Core Ultra 5 235H takes 5. That raw count, however, hides a more interesting story about what each chip does best.

The 235T dominates in compute-heavy workloads. Its biggest margins come from PassMark find prime numbers (33.1% ahead), floating point math (13.9% ahead), and integer math (13.5% ahead). These are classic CPU-bound tasks that stress the execution cores. The 235T also leads in every Cinebench test, from R15 through R23, in both single-core and multi-core variants, consistently by 2.5%. Physics simulation, multithreaded throughput, and data encryption also fall in its column.

The 235H, meanwhile, shows strength in a narrower set of tasks. It wins PassMark data compression by 2.3%, random string sorting by 2.3%, extended instructions by 0.6%, and single-thread performance by 0.5%. These wins are smaller in magnitude than the 235T's best victories. The mobile chip's advantages are real but modest, suggesting it has a slight edge in memory-latency-sensitive or string-handling workloads rather than raw arithmetic throughput.

What does this mean practically? The 235T is the choice when the workload is mathematical, scientific, or rendering-oriented. The 235H holds its own in data manipulation and single-threaded responsiveness. Neither chip is a runaway winner across the board, but the 235T's larger margins in its winning categories give it the higher average benchmark score. The database records an average score of 38561 for the 235T versus 37522 for the 235H, a difference of about 2.8%.

Architecture Differences

Both processors share the Arrow Lake architecture, the Core Ultra Series 2 generation, and a TSMC 3 nm process node. Both have 14 cores and 14 threads, meaning no hyperthreading on either part. The L1 and L2 caches are identical: 192 KB per core and 3 MB per core respectively. Both have dual-channel memory buses with 102.4 GB/s bandwidth, and neither supports ECC memory. Both are locked parts with no unlocked multiplier.

The divergences begin with packaging and platform. The 235T is a desktop chip on Intel Socket 1851, using the Arrow Lake-S die. The 235H is a mobile chip on Intel BGA 2049, using the Arrow Lake-H die. This is the most fundamental split: one is designed to sit in a desktop motherboard, the other to be soldered into a laptop.

The shared L3 cache differs significantly. The 235T has 24 MB of shared L3, while the 235H has 18 MB. That 6 MB difference likely contributes to the desktop part's stronger performance in cache-sensitive workloads. The process node is the same 3 nm TSMC fabrication, but the desktop part reports 17,800 million transistors on a 243 mm² die, while the database records no transistor count or die size for the mobile part.

Memory support also diverges. The 235T supports DDR5 only. The 235H supports both DDR5 and LPDDR5X, reflecting its mobile nature where low-power memory is common. PCIe connectivity shows a similar split: the 235T offers Gen 5 with 20 lanes (CPU only), while the 235H offers Gen 5 with just 8 lanes (CPU only). The desktop chip has more than double the PCIe lanes for expansion cards and peripherals.

Integrated graphics differ as well. The 235T carries Arc Xe-LPG Graphics with 24 execution units. The 235H carries Arc Graphics 140T. The database does not provide execution unit counts for the mobile part, so a direct graphics comparison is not possible from the recorded data.

Thermal design power tells the platform story. The 235T has a 65 W TDP, while the 235H has a 28 W TDP. The desktop chip is allowed to draw more power, which helps explain its higher performance in sustained workloads. The base clocks differ: the 235T runs at 2.20 GHz base, the 235H at 2.40 GHz base. Both boost to 5.00 GHz. The mobile chip starts from a higher base clock but has a much lower power envelope.

The release dates are close: the 235T came out on 2025-01-06, the 235H on 2025-01-12. Both are marked as Active in production status. The 235T has a launch MSRP of $247. The database records no launch MSRP for the 235H.

Head-to-Head Benchmarks

The Cinebench suite shows a consistent pattern. Across all six Cinebench tests, the 235T wins by exactly 2.5%. In R23 multi-core, the 235T scores 26232 against 25598 for the 235H. In R23 single-core, it is 3703 against 3613. The same 2.5% margin appears in R15 and R20, both single and multi-core variants. This uniformity suggests the desktop part's advantage is structural, likely tied to its higher power budget and larger L3 cache rather than any workload-specific behavior.

The PassMark suite tells a more varied story. The largest single gap is in find prime numbers, where the 235T scores 318 versus 239 for the 235H, a 33.1% advantage. This is an extreme outlier compared to the rest of the results. Floating point math shows a 13.9% lead for the 235T (106546 versus 93509), and integer math a 13.5% lead (84244 versus 74247). These are substantial margins that point to the desktop chip's superior sustained arithmetic throughput.

Physics simulation favors the 235T by 8.7% (2157 versus 1985). Multithreaded performance favors it by 2.7% (30918 versus 30091). Data encryption shows a smaller 1.5% lead (23457 versus 23121).

The 235H's wins are all below 2.5% in magnitude. Data compression goes to the mobile chip by 2.3% (301979 versus 295100). Random string sorting also goes to the 235H by 2.3% (36208 versus 35377). Extended instructions show a 0.6% edge (23354 versus 23212). Single-thread performance, recorded twice in the database under slightly different test names, gives the 235H a 0.5% advantage (4359 versus 4339).

The pattern is clear: the 235T wins by large margins in arithmetic and physics-heavy tasks, while the 235H wins by small margins in memory-oriented and single-threaded tasks. The desktop part's wins are bigger, which is why its average benchmark score sits higher. The 235H's wins are narrow enough that they barely move the overall average.

FAQ

Q: Which processor has the higher average benchmark score?

A: The Intel Core Ultra 5 235T has an average benchmark score of 38561, while the Intel Core Ultra 5 235H scores 37522. The 235T sits at the 86th percentile of all CPUs, the 235H at the 85th.

Q: How do their Cinebench R23 scores compare?

A: In Cinebench R23 multi-core, the 235T scores 26232 and the 235H scores 25598, a 2.5% lead for the desktop chip. In R23 single-core, the 235T scores 3703 and the 235H scores 3613, also a 2.5% lead.

Q: What is the biggest performance gap between the two?

A: The largest margin is in PassMark find prime numbers, where the 235T scores 318 and the 235H scores 239. That is a 33.1% advantage for the desktop part.

Q: Are there any benchmarks where the 235H wins?

A: Yes. The 235H wins PassMark data compression (301979 versus 295100), random string sorting (36208 versus 35377), extended instructions (23354 versus 23212), and single-thread performance (4359 versus 4339). All these wins are 2.3% or smaller.

Q: Do both processors have the same core and thread counts?

A: Yes. Both have 14 cores and 14 threads. They also share the same L1 cache (192 KB per core) and L2 cache (3 MB per core).

Q: What are the TDP differences?

A: The 235T has a 65 W TDP and is a desktop processor on Intel Socket 1851. The 235H has a 28 W TDP and is a mobile processor on Intel BGA 2049.

Specification Differences

The two processors differ in several recorded specifications. The 235T has a base clock of 2.20 GHz, while the 235H has a base clock of 2.40 GHz. Both share a 5.00 GHz boost clock. The TDP differs significantly: 65 W for the 235T versus 28 W for the 235H.

The L3 cache is larger on the desktop part: 24 MB shared versus 18 MB shared. The 235T uses Intel Socket 1851, the 235H uses Intel BGA 2049. The codenames reflect their platforms: Arrow Lake-S for the desktop, Arrow Lake-H for the mobile.

Memory support differs. The 235T supports DDR5. The 235H supports DDR5 and LPDDR5X. PCIe lanes also differ: the 235T has Gen 5 with 20 lanes (CPU only), the 235H has Gen 5 with 8 lanes (CPU only).

Integrated graphics are different. The 235T has Arc Xe-LPG Graphics with 24 execution units. The 235H has Arc Graphics 140T. The market segments differ: Desktop for the 235T, Mobile for the 235H.

The 235T reports 17,800 million transistors on a 243 mm² die. The database records no such figures for the 235H. The 235T has a launch MSRP of $247; the 235H has no recorded launch MSRP. The part numbers differ: SRQES for the 235T, SRQAP for the 235H. Release dates are 2025-01-06 for the 235T and 2025-01-12 for the 235H.

The Verdict

The data points to a straightforward conclusion for most workloads. The Intel Core Ultra 5 235T is the stronger processor. It wins 12 of 17 benchmark comparisons, holds the higher average score (38561 versus 37522), and achieves its wins by larger margins. The 33.1% lead in find prime numbers, the 13.9% lead in floating point math, and the 13.5% lead in integer math are not trivial gaps. Anyone running scientific calculations, financial modeling, physics simulations, or heavy multi-threaded rendering should favor the 235T.

The 235H is not without merit. Its wins in data compression, random string sorting, and single-thread performance suggest it handles data manipulation tasks competitively. The 0.5% single-thread advantage is small but consistent across both recorded single-thread tests. The mobile chip also offers LPDDR5X memory support and a much lower 28 W TDP, which matters for laptop designs where power and thermals are constrained. Its 8 PCIe Gen 5 lanes, while fewer than the desktop part's 20, are sufficient for mobile use cases.

The choice depends on platform. If the system is a desktop with Socket 1851, the 235T is the clear pick. Its higher TDP, larger L3 cache, more PCIe lanes, and better benchmark results all align. If the system is a laptop with BGA 2049, the 235H is the only option of the two. The 28 W TDP and LPDDR5X support are mobile-specific features that the desktop part cannot offer.

The nearest rival data reinforces the 235T's position. Its closest competitors in the database are the Intel Core i5-14500 (0.1% ahead), Intel Xeon w3-2525 (0.4% ahead), Intel Core 9 270H (0.6% ahead), and Intel Core Ultra 9 285H (0.7% ahead). The 235H's rivals include the Intel Core i9-13900HK (0.3% behind), Intel Core i5-13600K (0.4% ahead), Intel Core Ultra 5 225F (0.6% ahead), and AMD Ryzen AI 5 PRO 435 (0.6% behind). Both chips sit in competitive territory, but the 235T's rivals are generally closer to it in average score, and it edges several of them.

For a desktop builder, the 235T offers the better combination of cache, power headroom, and benchmark performance. For a laptop buyer, the 235H delivers competitive single-thread speed and respectable multi-core results within a 28 W envelope. The data does not support the 235H as a performance leader, but it does support it as a capable mobile processor that trades some arithmetic throughput for efficiency and platform flexibility.

DETAILED SPECIFICATIONS

SPECIFICATION
Ultra 5 235H
Ultra 5 235T
Core Specs
Cores
14
14 0.0%
Threads
14
14 0.0%
Base Clock (GHz)
2.4
2.2 -8.3%
Boost Clock (GHz)
5
5 0.0%
Frequency (GHz)
2.4
2.2 -8.3%
Turbo Clock (GHz)
5
5 0.0%
Multiplier
24
22 -8.3%
SMP CPUs
1
1 0.0%
Cache
L1 Cache
192 KB (per core)
192 KB (per core)
L2 Cache
3 MB (per core)
3 MB (per core)
L3 Cache
18 MB (shared)
24 MB (shared)
Power
TDP (W)
28
65 +132.1%
PL1
28 W
35 W
PL2
60 W
114 W
Architecture
Architecture
Arrow Lake
Arrow Lake
Codename
Arrow Lake-H
Arrow Lake-S
Generation
Ultra 5 (Arrow Lake-H)
Ultra 5 (Arrow Lake)
Process Size
3 nm
3 nm
Transistors
—
17,800 million
Die Size
—
243 mm²
Foundry
TSMC
TSMC
Memory
Memory Support
DDR5, LPDDR5X
DDR5
Memory Bus
Dual-channel
Dual-channel
Memory Bandwidth
102.4 GB/s
102.4 GB/s
ECC Memory
No
No
Platform
Socket
Intel BGA 2049
Intel Socket 1851
Chipsets
WM880, HM870
Z890, B860, W880, Q870, H810
PCIe
Gen 5, 8 Lanes(CPU only)
Gen 5, 20 Lanes(CPU only)
Intel Hybrid
Hybrid Cores
P-Cores: 4 E-Cores: 10
P-Cores: 6 E-Cores: 8
E-Core Frequency
1800 MHz up to 4.4 GHz
1600 MHz up to 4.4 GHz
P-Core Turbo
—
4.8 GHz
LP E-Cores
2
—
AI/NPU
NPU
Yes / 13 TOPS
—
Graphics
Integrated Graphics
Arc Graphics 140T
Arc Xe-LPG Graphics 24EU
Other
Market
Mobile
Desktop
Production Status
Active
Active
Launch Price
—
$247
Part Number
SRQAP
SRQES
Package
FC-BGA
FC-LGA18W
Tj Max
110°C
105°C
View Core Ultra 5 235H Details View Core Ultra 5 235T Details