AMD Ryzen AI 7 450G vs Intel Core Ultra 7 268V Comparison
AMD Ryzen AI 7 450G
Core Ultra 7 268V
PERFORMANCE BENCHMARKS
Analysis: AMD Ryzen AI 7 450G vs Intel Core Ultra 7 268V
The AMD Ryzen AI 7 450G and Intel Core Ultra 7 268V are both 8-core processors, but they target different segments and produce very different benchmark profiles. The AMD part is a desktop AM5 chip with 16 threads, a 5.10 GHz boost, and a 65 W TDP; the Intel part is a mobile BGA chip with 8 threads, a 5.00 GHz boost, and a 17 W TDP. In the head-to-head PassMark set, the AMD wins 9 of 11 tests and the Intel wins 2, with the AMD sitting at the 93rd percentile of all CPUs and the Intel at the 74th.
Where Each One Wins
AMD's wins cluster around throughput and data manipulation. It leads by 122% in data compression, 402,831 versus 181,443, and by 133.7% in integer math, 99,732 versus 42,669. Random string sorting goes AMD's way by 90.3%, 42,663 versus 22,416, and extended instructions by 84.2%, 28,223 versus 15,323. Those four workloads define the AMD's strength in compression, sorting, and integer-heavy processing. The multithread score extends that pattern, 30,422 versus 19,297, a 57.7% margin. Data encryption also favors AMD, 18,937 versus 13,779, a 37.4% lead. Floating point math is closer, 63,683 versus 57,628, a 10.5% edge. Single-thread performance is the narrowest AMD win, 4,309 versus 4,051, a 6.4% margin.
Intel's two wins are isolated. The prime-number test goes to Intel, 192 versus 87, which means AMD trails by 54.7% in that workload. Physics also favors Intel, 1,617 versus 1,440, a 10.9% AMD deficit. The data does not show Intel winning any of the broad throughput categories; its recorded wins are in prime-number finding and physics.
Architecture Differences
Both chips are built by TSMC, but on different nodes: AMD uses 4 nm, Intel uses 3 nm. The core layouts differ as well. AMD implements 8 cores and 16 threads under the Ryzen AI 400 generation label, which includes Zen 5 and Zen 5c. Intel implements 8 cores and 8 threads under the Ultra 7 Lunar Lake generation. AMD's cache is 80 KB L1 per core, 1 MB L2 per core, and 8 MB L3. Intel's cache is 192 KB L1 per core, 2.5 MB L2 per core, and 12 MB shared L3. That gives Intel more cache per core and a larger shared L3, while AMD processes twice as many threads.
Power and platform differences are large. AMD has a 65 W TDP, an AM5 socket, and an unlocked multiplier. Intel has a 17 W TDP, a BGA 2833 socket, and a locked multiplier. AMD is classified as desktop, Intel as mobile. AMD supports DDR5 and LPDDR5X with dual-channel memory and 89.6 GB/s bandwidth, plus ECC. Intel's memory support is listed as unknown and depends on motherboard, with dual-channel but no recorded bandwidth and no ECC. AMD exposes PCIe Gen 4 with 12 CPU lanes; Intel exposes PCIe Gen 5 with 4 CPU lanes. The integrated GPUs differ: Radeon 860M on AMD, Arc 140V on Intel. AMD's die size is recorded at 195 mm²; Intel's die size is not recorded. Release dates also differ: AMD 2026-02-28, Intel 2024-09-23.
Head-to-Head Benchmarks
Across the 11 recorded head-to-head entries, AMD wins 9 and Intel wins 2. The largest margin is integer math: AMD 99,732 versus Intel 42,669, a 133.7% advantage. Data compression follows at 122%, 402,831 versus 181,443. Random string sorting gives AMD a 90.3% win, 42,663 versus 22,416, and extended instructions give 84.2%, 28,223 versus 15,323. Multithread is 57.7%, 30,422 versus 19,297. Data encryption is 37.4%, 18,937 versus 13,779. Floating point math is 10.5%, 63,683 versus 57,628. Single-thread is 6.4%, 4,309 versus 4,051.
Intel's wins are fewer but clear. Prime-number finding: Intel 192, AMD 87, AMD trails by 54.7%. Physics: Intel 1,617, AMD 1,440, AMD trails by 10.9%. The Intel's own benchmark set also records Cinebench R23 multi-core 10,653, single-core 1,921, and Geekbench multi-core 9,963, single-core 2,270.
The Verdict
The database paints a clear split. For multithreaded desktop processing, compression, integer math, sorting, encryption, and even single-thread speed, the AMD Ryzen AI 7 450G is the stronger recorded part. It holds a 57.7% multithread lead and a 6.4% single-thread lead. Its average score of 63,331 places it at the 93rd percentile. The nearest rivals in the database are AMD Ryzen AI Max PRO 390 at 63,765, 0.7% higher, Intel Core Ultra 7 265HX at 63,173, 0.2% lower, Intel Core i7-13790F at 63,080, 0.4% lower, and AMD Ryzen AI Embedded P185 at 62,839, 0.8% lower.
Intel's case is narrower. The Core Ultra 7 268V averages 20,897, at the 74th percentile. Its nearest rivals are AMD Ryzen 5 PRO 4655GE at 20,900, 0% difference, Intel Core i5-12600T at 20,917, 0.1% higher, AMD EPYC 7J13 at 20,845, 0.2% lower, and Intel Core Ultra 5 125U at 20,826, 0.3% lower. Its two head-to-head wins are prime-number finding and physics, and it operates at a 17 W TDP versus 65 W, which aligns with its mobile segment. For a desktop build running data-heavy and multithreaded workloads, the recorded data points to AMD. For a low-power mobile system where those two Intel benchmark wins and the 17 W envelope are priorities, the Intel is the supported choice.
FAQ
Q: How do the core and thread counts compare?
A: Both processors have 8 cores. The AMD Ryzen AI 7 450G has 16 threads, while the Intel Core Ultra 7 268V has 8 threads.
Q: Which processor wins the most head-to-head benchmarks?
A: AMD wins 9 of the 11 shared PassMark tests. Intel wins the remaining 2: prime-number finding and physics.
Q: What is the single-thread performance gap?
A: AMD scores 4,309 versus Intel's 4,051 in PassMark single-thread, a 6.4% lead for AMD.
Q: What are the TDP and platform differences?
A: AMD is a 65 W desktop part on Socket AM5 with an unlocked multiplier. Intel is a 17 W mobile part on BGA 2833 with a locked multiplier.
Q: Which processor supports ECC memory?
A: AMD lists ECC as enabled. Intel lists ECC as false.
Q: What are the process nodes?
A: AMD uses TSMC 4 nm, Intel uses TSMC 3 nm.
Specification Differences
| Field | AMD Ryzen AI 7 450G | Intel Core Ultra 7 268V |
|---|---|---|
| Series | Not recorded | Core Ultra Series 2 |
| Manufacturer | AMD | Intel |
| Threads | 16 | 8 |
| Base clock | 2.00 GHz | 2.20 GHz |
| Boost clock | 5.10 GHz | 5.00 GHz |
| TDP | 65 W | 17 W |
| Socket | AMD Socket AM5 | Intel BGA 2833 |
| Architecture | Not recorded | Lunar Lake |
| Codename | Gorgon Point | Lunar Lake |
| Generation | Ryzen AI 400 (Zen 5 / Zen 5c) | Ultra 7 (Lunar Lake) |
| Process node | 4 nm | 3 nm |
| Die size | 195 mm² | Not recorded |
| L1 cache | 80 KB per core | 192 KB per core |
| L2 cache | 1 MB per core | 2.5 MB per core |
| L3 cache | 8 MB | 12 MB shared |
| Memory support | DDR5, LPDDR5X | Unknown, depends on motherboard |
| Memory bandwidth | 89.6 GB/s | Not recorded |
| ECC memory | Yes | No |
| PCIe | Gen 4, 12 lanes (CPU only) | Gen 5, 4 lanes (CPU only) |
| Integrated graphics | Radeon 860M | Arc 140V |
| Market segment | Desktop | Mobile |
| Release date | 2026-02-28 | 2024-09-23 |
| Multiplier unlocked | Yes | No |
| Part number | 100-000001917 | SRPMLSRPMX |