Apple's M5 Max Outperforms 96-Core Threadripper in Geekbench, But GPU Falls Short

Apple's latest M5 Max processor has delivered a surprising performance breakthrough, outperforming AMD's flagship 96-core Threadripper Pro 9995WX in both single-thread and multi-thread workloads according to recent Geekbench 6 benchmark results. The 18-core M5 Max achieved 4,353 points in single-thread and 29,644 points in multi-thread performance, surpassing the Threadripper's typical scores of around 26,000 points in multi-threaded tests.

This achievement is particularly noteworthy given the massive core count disparity - Apple's chip packs just 18 cores compared to AMD's 96-core behemoth. The M5 Max's victory in multi-threaded workloads highlights how benchmark characteristics can favor different architectures. Geekbench 6's multi-thread test is designed as a brief, bursty workload that mimics consumer tasks like archive compression and image editing. This short runtime and the benchmark's limited thread scaling - which efficiently utilizes only 8-32 threads for many subtests - creates an environment where Apple's fewer but faster cores can outperform AMD's massive core count.

However, there's a caveat to this comparison. One Threadripper Pro 9995WX sample achieved 30,170 points in multi-thread, slightly edging out the M5 Max. Additionally, Geekbench 6 is a synthetic benchmark that may not reflect real-world application performance, particularly for workloads that can fully utilize hundreds of cores.

The M5 Max's impressive performance stems from several architectural improvements. The chip features six 'super' performance cores with enhanced front-end bandwidth, improved branch prediction, and a new cache hierarchy. These are complemented by 12 performance cores, up from 16 total cores (12P + 4E) in the M4 Max. The memory subsystem also saw upgrades, with LPDDR5X-9600 memory providing 614 GB/s of bandwidth through a 512-bit interface - a 12% increase over the M4 Max's 546 GB/s.

When compared to other high-end processors, the M5 Max's performance is equally impressive. It outpaces Intel's Core Ultra 9 285K (21,014 points), AMD's Ryzen 9 9950X3D (24,057 points), and even Intel's Xeon W9-3595X (24,206 points) in multi-thread workloads. The single-thread performance of 4,353 points also exceeds all these competitors.

GPU Performance: Strong for Integrated, Weak for Discrete

While the CPU performance story is one of triumph, the M5 Max's integrated GPU tells a more modest tale. The new GPU, based on a PowerVR-derived microarchitecture, scores 232,718 points in Geekbench 6's GPU compute benchmark using the Metal API. This represents an improvement over the M4 Max's 204,453 points, but the gains are incremental rather than revolutionary.

In the integrated GPU space, the M5 Max easily outperforms AMD's Ryzen AI Max+ 395 (133,447 points). However, when compared to discrete graphics cards, the limitations become apparent. The M5 Max trails Nvidia's GeForce RTX 5070 (207,061 points, Vulkan) and is significantly behind the RTX 5070 Ti (253,890 points) and RTX 5090 (376,858 points).

Despite these limitations, Apple's achievement in creating an integrated GPU that can compete with mid-range discrete graphics cards represents a significant milestone. The M5 Max's GPU demonstrates strong performance in specific workloads, particularly in particle physics simulations where it achieves 23,215 FPS compared to the RTX 5090's 47,086 FPS.

Market Implications

The M5 Max's performance breakthrough has several important implications for the processor market. First, it demonstrates that Apple's chip design philosophy - focusing on per-core performance and efficiency rather than raw core count - continues to yield dividends. This approach appears particularly effective in the types of workloads that dominate consumer and professional computing.

Second, the results highlight the limitations of using synthetic benchmarks for cross-platform comparisons, especially when dealing with vastly different core counts and architectures. The bursty nature of Geekbench 6's multi-thread test particularly favors Apple's architecture while potentially underutilizing AMD's massive core count.

For creative professionals and developers who work in Apple's ecosystem, the M5 Max offers compelling performance that can handle demanding workloads without requiring the massive cooling solutions needed by high-core-count desktop processors. The combination of strong single-thread performance and competitive multi-thread capabilities makes it a versatile choice for a wide range of applications.

However, users requiring maximum GPU performance for tasks like 3D rendering, scientific computing, or machine learning may still find dedicated graphics cards more suitable, as even Apple's best integrated solution cannot match the performance of high-end discrete GPUs.

The M5 Max represents another step in Apple's chip design evolution, showcasing how focused architectural improvements can yield performance gains that outpace competitors with significantly more resources. As Apple continues to refine its processor designs, the gap between its integrated solutions and traditional PC architectures may continue to narrow in certain workloads while remaining substantial in others.