Intel Core Ultra 9 285T vs Intel Xeon 634 Comparison

Where Each One Wins

The Intel Xeon 634 and Intel Core Ultra 9 285T occupy distinctly different positions in the performance landscape. The recorded data shows the Core Ultra 9 285T wins 15 of the 17 head-to-head benchmark comparisons, while the Xeon 634 claims only 2. However, those two wins are decisive and point toward a clear specialization.

The Xeon 634 dominates in data compression and extended instruction workloads. Its PassMark data compression score of 477924 is 24.4% ahead of the Core Ultra 9 285T's 384140. Similarly, in extended instructions, the Xeon scores 38320 versus 27477, a 39.5% advantage. These are enterprise-oriented workloads where the Xeon's server heritage shows.

The Core Ultra 9 285T takes everything else. It wins all six Cinebench tests, every PassMark math and physics test, data encryption, multithread, random string sorting, and single-thread performance. The margin varies from a narrow 1.4% in random string sorting to a substantial 43.2% in find prime numbers.

For general desktop use, productivity, and single-threaded responsiveness, the Core Ultra 9 285T is the clear choice. For server data pipelines, compression-heavy tasks, and workloads that leverage extended instruction sets, the Xeon 634 holds the advantage. The average benchmark score reflects this split: the Xeon 634 averages 52974 across all tests, slightly ahead of the Core Ultra 9 285T's 51310, despite losing most individual comparisons, because its two big wins skew the average upward.

Architecture Differences

The two processors come from different Intel design lineages. The Xeon 634 uses the Granite Rapids architecture on a 5 nm process node fabricated by Intel, with a die size of 598 mm². It is a server and workstation part with a 12-core, 24-thread configuration, a base clock of 2.70 GHz, and a boost clock of 4.60 GHz. Its cache hierarchy includes 112 KB of L1 per core, 2 MB of L2 per core, and 48 MB of shared L3.

The Core Ultra 9 285T is built on Arrow Lake, a 3 nm process node from TSMC, with a significantly smaller die at 243 mm². It packs 24 cores and 24 threads, meaning no hyperthreading, with a base clock of 1.40 GHz and a boost clock of 5.40 GHz. Its cache includes 192 KB of L1 per core, 3 MB of L2 per core, and 36 MB of shared L3. The transistor count is listed at 17,800 million for the Core Ultra 9 285T, while the Xeon 634 does not have a recorded transistor figure.

Memory support differs substantially. Both support DDR5, but the Xeon 634 runs quad-channel with a memory bandwidth of 204.8 GB/s, while the Core Ultra 9 285T runs dual-channel at 102.4 GB/s. This is a critical distinction for memory-bound server workloads. Both support ECC memory, which is notable for a desktop part like the Core Ultra 9 285T.

PCIe connectivity also diverges. The Xeon 634 provides Gen 5 with 80 lanes, while the Core Ultra 9 285T provides Gen 5 with 20 lanes. Integrated graphics are absent on the Xeon 634, but the Core Ultra 9 285T includes Arc Xe-LPG Graphics with 64 execution units. The Xeon 634 has an unlocked multiplier, while the Core Ultra 9 285T does not.

Head-to-Head Benchmarks

The Cinebench results show a consistent pattern. Across R15, R20, and R23, both multicore and singlecore, the Core Ultra 9 285T wins every test by exactly 4.8%. In R15 multicore, the Core Ultra scores 3384 against the Xeon's 3220. In R15 singlecore, it is 477 versus 454. R20 multicore shows 14100 against 13419, and R20 singlecore shows 1990 against 1894. R23 multicore yields 33573 versus 31950, and R23 singlecore yields 4739 versus 4510.

The PassMark suite tells a more varied story. The Xeon 634's two wins are decisive. Data compression at 477924 versus 384140 represents a 24.4% lead, and extended instructions at 38320 versus 27477 represents a 39.5% lead. These are the only benchmarks where the Xeon comes out ahead, and both are substantial margins.

The Core Ultra 9 285T's wins include some large gaps. In find prime numbers, it scores 345 versus 196, a 43.2% advantage. Floating point math shows 137923 versus 93564, a 32.2% lead. Data encryption comes in at 32061 versus 23451, a 26.9% margin. Physics shows 2842 versus 2250, a 20.8% lead. Single-thread performance is 4576 versus 3567, a 22% difference.

Smaller margins appear in other tests. Integer math is 132433 versus 117664, an 11.2% lead. Multithread is 39931 versus 37589, a 5.9% margin. Random string sorting is nearly identical at 47695 versus 47016, just 1.4% apart. The closest competition is in random string sorting, while the widest gaps favor the Core Ultra 9 285T in prime number finding and the Xeon 634 in extended instructions.

Specification Differences

The two processors differ on nearly every specification level. Core counts are 12 for the Xeon 634 and 24 for the Core Ultra 9 285T, though both have 24 threads. Base clocks are 2.70 GHz for the Xeon and 1.40 GHz for the Core Ultra, while boost clocks are 4.60 GHz and 5.40 GHz respectively. The Xeon has a TDP of 150 watts, while the Core Ultra 9 285T draws only 35 watts. This power difference is stark and has implications for thermal management and operating cost, though the performance per watt data is not directly recorded.

Sockets are incompatible: the Xeon 634 uses Intel Socket 4710, the Core Ultra 9 285T uses Intel Socket 1851. Process nodes differ at 5 nm for the Xeon and 3 nm for the Core Ultra, with different foundries: Intel for the former, TSMC for the latter. Die sizes are 598 mm² versus 243 mm².

Cache configurations differ in size and distribution. The Xeon has 112 KB L1 per core, 2 MB L2 per core, and 48 MB shared L3. The Core Ultra has 192 KB L1 per core, 3 MB L2 per core, and 36 MB shared L3. Memory channels are quad-channel for the Xeon and dual-channel for the Core Ultra, with corresponding bandwidth figures of 204.8 GB/s and 102.4 GB/s.

PCIe lane counts are 80 for the Xeon versus 20 for the Core Ultra. Integrated graphics are absent on the Xeon but present on the Core Ultra as Arc Xe-LPG Graphics 64EU. The market segments differ: Server/Workstation versus Desktop. Release dates are February 2026 for the Xeon and January 2025 for the Core Ultra. The launch MSRP is $499 for the Xeon 634 and $549 for the Core Ultra 9 285T. Multiplier unlocking is available on the Xeon but not on the Core Ultra. Part numbers are SA2DL and SRQD3 respectively.

FAQ

Q: Which processor has more cores?

A: The Intel Core Ultra 9 285T has 24 cores, while the Intel Xeon 634 has 12 cores. Both have 24 threads, since the Xeon uses hyperthreading and the Core Ultra does not.

Q: Which processor is faster in single-threaded performance?

A: The Core Ultra 9 285T is faster. In PassMark single thread, it scores 4576 versus 3567 for the Xeon 634, a 22% advantage. In Cinebench R23 singlecore, it scores 4739 versus 4510, a 4.8% lead.

Q: What are the biggest performance gaps between the two processors?

A: The largest gap favors the Core Ultra 9 285T in PassMark find prime numbers, where it leads by 43.2% (345 versus 196). The largest gap favoring the Xeon 634 is in PassMark extended instructions, where it leads by 39.5% (38320 versus 27477).

Q: Do both processors support ECC memory?

A: Yes, both the Intel Xeon 634 and the Intel Core Ultra 9 285T support ECC memory. This is unusual for a desktop part like the Core Ultra 9 285T.

Q: Which processor has higher memory bandwidth?

A: The Xeon 634 has double the memory bandwidth of the Core Ultra 9 285T. The Xeon operates at 204.8 GB/s with quad-channel memory, while the Core Ultra operates at 102.4 GB/s with dual-channel memory.

Q: How do the average benchmark scores compare?

A: The Xeon 634 has an average benchmark score of 52974, while the Core Ultra 9 285T has an average of 51310. Both processors rank at the 91st percentile among all CPUs in the database. The Xeon's higher average comes from its large wins in data compression and extended instructions, despite losing most individual comparisons.