The Intel Core Ultra 7 265H is a mobile high-end CPU for larger notebooks based on the Arrow Lake architecture. It offers 16 cores consisting of 6 fast performance cores (Lion Cove) with up to 5.3 GHz and 8 smaller efficiency cores (Skymont) with up to 4.5 GHz clock speed and two additional "low power" efficiency cores with up to 2.5 GHz (Skymont). The CPU has access to 24 MB of cache and is specified with a TDP of 28 watt (PL1, 115W PL2).
The SoC integrates a small dedicated NPU called AI Boost with 13 TOPS (Int8) and optionally supports vPro Enterprise. The integrated memory controller supports up to 192 GB LPDDR5/x-8400 or DDR5-6400 (dual channel). The integrated GPU (iGPU) is an Intel Arc 140T graphics card which offers eight Xe cores with up to 2.3 GHz.
Performance - High End
Compared to the fastest model of the Arrow Lake-H series (Core Ultra 9 285H), the Core Ultra 7 as only slightly slower due to a reduced maximum Turbo frequency of 100 MHz. Thanks to the improved IPC of the cores (especially the efficiency cores), the CPU should reach the level of the old Raptor Lake Core Core 9 270H (6 + 8 cores, max 5.8 / 4.1 GHz). This makes the CPU ideal for very demanding tasks such as high-end gaming and content creation.
Chiplet design
Like the desktop and HX chips, the Arrow Lake-H chips are based on individual chiplets that are placed on a 22nm base tile from Intel using Foveros 3D packaging. The CPU part comes from TSMC using the modern N3B (3nm) process. As with HX, the GPU also is produced at TSMC and is manufactured using the N5P process. SoC and I/O Tile are also from TSMC but in the older N6 process.
The Intel Core Ultra 7 265U is an upper mid-range CPU with 12 cores from the Arrow Lake U series, which was introduced at the beginning of 2025.The SoC is based on a tile/chiplet design. The chip is similar to the Meteor Lake U series, but the compute part is manufactured using the modern Intel 3 process and therefore offers higher clock rates.
The smaller compute tile (compared to the H processors) offers 2 performance cores (P cores, Redwood Cove architecture, SMT) and 8 efficiency cores (E cores, Crestmont architecture). The P-cores support hyperthreading (therefore 14 threads together) and clock up to 5.3 GHz. The E-cores clock at a maximum of 4.2 GHz. In addition, there are now two more low-power efficiency cores on the low-power island, so the compute tile can be completely deactivated when idle and during simple tasks in order to save power.
Performance
Thanks to the significantly higher clock rates, the Core Ultra 7 265U should be significantly faster than the old Core Ultra 7 165U (12 cores, max 4.9 / 3.8 GHz, 12 MB cache). Due to the 12 cores, the multi-thread performance is very good, but can be limited by the low TDP (especially under long load).
Graphics unit
The Core Ultra 7 265U offers a new Arc-based iGPU called Intel Graphics with 4 Xe cores and 64 EUs that reach a maximum of 2.1 GHz. However, the performance is clearly below the new Arc iGPU in the H series.
Features
Arrow Lake-U has integrated WiFi 6E (WiFi 7 optional) and Thunderbolt 4 (4x). The integrated memory controller now supports LPDDR5/x-8400 or DDR5-6400 (max. 128 GB). A dedicated NPU (2x Gen3 Neural Compute Engines) is now used. The Xe Media Engine supports MPEG-2, AVC, VC-1 Decode, JPEG, VP8 Decode, VP9, HEVC and AV1 Decode up to 8K 10-bit HDR. The chip now supports PCIe-5.0 (x8 for GPU) and PCIe-4.0 (three x12 for SSDs).
Power consumption
The Intel Core Ultra 7 265U is specified with 15 watts TDP (base) and 57 watts (PL2) max. turbo power. Meteor Lake consists of five individual chips, whereby the processor part is manufactured in the new Intel 3 process (5nm). The graphics unit is produced at TSMC in N5 and the SoC and I/O tile in the older N6 process. These 4 chips are then applied to the 22nm base tile using the Foveros process.
Average Benchmarks Intel Core Ultra 7 265H → 100%n=25
Average Benchmarks Intel Core Ultra 7 265U → 68%n=25
- Range of benchmark values for this graphics card - Average benchmark values for this graphics card * Smaller numbers mean a higher performance 1 This benchmark is not used for the average calculation
v1.35
log 11. 02:14:14
#0 ran 0s before starting gpusingle class +0s ... 0s
#1 checking url part for id 18901 +0s ... 0s
#2 checking url part for id 18739 +0s ... 0s
#3 redirected to Ajax server, took 1762823654s time from redirect:0 +0s ... 0s
#4 did not recreate cache, as it is less than 5 days old! Created at Mon, 10 Nov 2025 05:21:03 +0100 +0s ... 0s
#5 composed specs +0.006s ... 0.006s
#6 did output specs +0s ... 0.006s
#7 getting avg benchmarks for device 18901 +0.004s ... 0.01s
#8 got single benchmarks 18901 +0.005s ... 0.016s
#9 getting avg benchmarks for device 18739 +0.004s ... 0.019s
#10 got single benchmarks 18739 +0.004s ... 0.023s
#11 got avg benchmarks for devices +0s ... 0.023s
#12 min, max, avg, median took s +0.026s ... 0.049s