Intel Core Ultra 5 115U vs Intel Core Ultra 5 134U Comparison

FAQ

Q: Which processor has more cores and threads?

A: The Intel Core Ultra 5 134U has 12 cores and 14 threads, while the Intel Core Ultra 5 115U has 8 cores and 10 threads.

Q: What is the biggest performance gap in the head-to-head benchmarks?

A: The largest delta is in Cinebench R23 multicore, where the Core Ultra 5 115U scores 10855 versus 5572 for the Core Ultra 5 134U, a 48.7% advantage.

Q: Which CPU wins the most head-to-head tests?

A: The Core Ultra 5 134U wins 9 of the 17 recorded head-to-head benchmarks, while the Core Ultra 5 115U wins 8.

Q: How do their single-thread scores compare?

A: The Core Ultra 5 134U leads in Cinebench R15 singlecore by 43.5% (221 vs 154) and in Cinebench R20 singlecore by 3.6% (666 vs 643), but the gap narrows to just 0.2% in PassMark single-thread tests (3063 vs 3057).

Q: What is the difference in shared L3 cache?

A: The Core Ultra 5 134U has 12 MB of shared L3 cache, while the Core Ultra 5 115U has 10 MB.

Q: Do both processors use the same socket?

A: No, the Core Ultra 5 134U uses Intel BGA 2551, while the Core Ultra 5 115U uses Intel BGA 2049.

Architecture Differences

Both processors belong to the Core Ultra Series 1 family and share the Meteor Lake architecture, codenamed Meteor Lake, built on Intel's 7 nm process node. The two chips diverge in core configuration, cache hierarchy, and integrated graphics.

The Core Ultra 5 134U packs 12 cores and 14 threads, while the Core Ultra 5 115U scales down to 8 cores and 10 threads. This difference in core count explains why the 134U posts higher results in several integer-heavy workloads, despite the 115U's higher base clock of 1500 MHz versus 700 MHz. The boost clocks are closer: 4.40 GHz for the 134U and 4.20 GHz for the 115U.

Cache layouts differ as well. Both chips feature 112 KB of L1 cache per core and 2 MB of L2 cache per core, but the shared L3 cache is 12 MB on the 134U versus 10 MB on the 115U. The larger L3 allocation on the 134U gives it more headroom for workloads that repeatedly access shared data sets.

The integrated graphics solutions are not identical. The Core Ultra 5 134U carries an Arc Xe-LPG 64EU GPU, while the Core Ultra 5 115U steps down to an Arc Xe-LPG 48EU GPU. That difference in execution units may matter for light graphics tasks, though the benchmark data in this comparison focuses on CPU workloads.

Memory support is listed as DDR5 with capacity depending on motherboard for both parts, and both use a dual-channel memory bus. The 115U records a memory bandwidth figure of 89.6 GB/s, while no bandwidth number is listed for the 134U in the database. PCIe lane counts differ: the 134U provides Gen 4 with 8 lanes (CPU only), and the 115U provides Gen 4 with 12 lanes (CPU only), so the 115U has more CPU-attached PCIe connectivity available.

The thermal design envelope also separates the two. The 134U has a TDP of 9 watts, while the 115U has a TDP of 15 watts. This means the 115U is allowed to draw more power, which is notable because it still wins the multicore Cinebench R23 test decisively despite having fewer cores. The 134U is the more power-efficient part on paper, but the 115U leverages its higher power budget in sustained multithreaded workloads.

Both chips use the same production status, are marked as Active, and share a release date of December 2023. Neither has an unlocked multiplier. The 134U has a launch MSRP of $332; the 115U has no launch MSRP recorded in the database.

Head-to-Head Benchmarks

The benchmark results paint a split picture: the Core Ultra 5 134U dominates single-thread and integer-oriented tests, while the Core Ultra 5 115U takes the lead in several multicore and memory-sensitive workloads.

Starting with Cinebench, the 115U wins Cinebench R15 multicore with a score of 1094 versus 894, an 18.3% margin. It also wins Cinebench R23 multicore by a massive 48.7%, scoring 10855 versus 5572. Interestingly, the 134U wins Cinebench R20 multicore by a narrow 3.6% margin (4725 vs 4559). This inconsistency across Cinebench versions suggests that workload scaling and power behavior differ significantly between the two chips under different test conditions.

The single-core Cinebench results favor the 134U across the board. In Cinebench R15 singlecore, the 134U scores 221 versus 154, a 43.5% advantage. In Cinebench R20 singlecore, it leads 666 to 643, a 3.6% margin. In Cinebench R23 singlecore, the 134U wins 1576 to 1532, a 2.9% edge. These results align with the 134U's higher boost clock of 4.40 GHz versus 4.20 GHz.

PassMark tests show a more even distribution. The 115U wins data compression (141945 vs 124771, a 12.1% edge), data encryption (8606 vs 8077, a 6.1% margin), extended instructions (8620 vs 7002, an 18.8% lead), multithread (13024 vs 12464, a 4.3% edge), physics (898 vs 883, a 1.7% margin), and random string sorting (15124 vs 13702, a 9.4% advantage). The 134U counters with wins in find prime numbers (56 vs 53, a 5.7% margin), floating point math (32425 vs 30538, a 6.2% edge), and integer math (51306 vs 41043, a substantial 25% lead). In PassMark single-thread and singlethread tests, the 134U wins by a razor-thin 0.2% margin (3063 vs 3057).

The overall win tally is 9 for the 134U and 8 for the 115U, but the magnitude of the wins matters more than the count. The 115U's 48.7% blowout in Cinebench R23 multicore is the single largest delta in the comparison, and its extended instructions lead of 18.8% is also substantial. The 134U's 25% integer math win and 43.5% Cinebench R15 singlecore win are equally emphatic, but the 115U's multicore dominance suggests it is the stronger choice for heavily threaded production workloads.

Specification Differences

The two processors differ in several key specification fields. The Core Ultra 5 134U has 12 cores and 14 threads, while the Core Ultra 5 115U has 8 cores and 10 threads. Base clocks are 700 MHz for the 134U and 1500 MHz for the 115U, while boost clocks are 4.40 GHz and 4.20 GHz, respectively. The TDP is 9 watts for the 134U and 15 watts for the 115U.

Socket compatibility differs: the 134U uses Intel BGA 2551, and the 115U uses Intel BGA 2049. Shared L3 cache is 12 MB on the 134U versus 10 MB on the 115U. The integrated GPU is Arc Xe-LPG 64EU on the 134U and Arc Xe-LPG 48EU on the 115U. PCIe support is Gen 4 with 8 lanes on the 134U and Gen 4 with 12 lanes on the 115U. The 115U lists a memory bandwidth of 89.6 GB/s, while the 134U does not have a recorded bandwidth value. The 134U has a launch MSRP of $332; the 115U has no launch MSRP recorded.

The part numbers also differ: SRN84 for the 134U and SRN6F for the 115U. Both processors share the same architecture, codename, generation, process node, foundry, L1 cache per core, L2 cache per core, memory support type, memory bus width, ECC support status, market segment, production status, release date, and unlocked multiplier state.

Where Each One Wins

The Core Ultra 5 134U is the better fit for single-threaded and latency-sensitive applications. Its 43.5% lead in Cinebench R15 singlecore and consistent, if smaller, wins in Cinebench R20 and R23 singlecore point to a stronger per-core ceiling. The 134U also wins integer math by 25%, which makes it attractive for workloads that rely heavily on integer operations, such as scripting, parsing, and certain database operations. Floating point math also favors the 134U by 6.2%, and it edges out the 115U in find prime numbers by 5.7%. The 134U's 12 MB L3 cache, larger than the 115U's 10 MB, may contribute to these wins in cache-sensitive tasks.

The Core Ultra 5 115U is the stronger choice for multithreaded and throughput-oriented workloads. Its 48.7% victory in Cinebench R23 multicore is the standout result, and it also wins Cinebench R15 multicore by 18.3%. The 115U's higher TDP of 15 watts, compared to 9 watts for the 134U, allows it to sustain higher power draw during extended multithreaded runs, which likely explains its multicore dominance despite having fewer cores. The 115U also wins data compression by 12.1%, extended instructions by 18.8%, and random string sorting by 9.4%, all of which are memory- or bandwidth-sensitive operations. Its 89.6 GB/s memory bandwidth figure, recorded in the database, supports this profile. The 115U also leads in data encryption by 6.1%, multithread by 4.3%, and physics by 1.7%.

For general productivity, the 115U's PassMark multithread win (13024 vs 12464) and its average benchmark score of 16753 versus 15910 for the 134U indicate that the 115U is the more balanced performer across the full test suite. The 134U, however, holds a higher percentile rank context: both chips sit at the 70th percentile against all CPUs, so neither is a dramatic outlier.

In summary, the 134U suits users who prioritize single-core responsiveness and integer-heavy tasks, while the 115U fits users who need multicore throughput, data compression speed, and extended instruction performance. The 134U's 9-watt TDP also positions it for more power-constrained designs, while the 115U's 15-watt TDP suggests it belongs in chassis with more thermal headroom. The choice hinges on workload mix: single-threaded and integer-bound work favors the 134U, multithreaded and bandwidth-heavy work favors the 115U.