Apple M5 Ultra vs Apple M3 Ultra

The M5 is the newer part: Apple M5, launched in 2026, against the M3 Ultra's Apple M3 from 2025. Newer architectures typically add lower-precision formats and better throughput per watt, so check the precision rows your workload actually uses.

Dense throughput for the M5 against the M3 Ultra: FP32 33 vs 33 TFLOPS. Sparse figures, where the vendor publishes them, appear under each dense number in the performance table.

Memory is 512 GB against 512 GB, fed at 1.2 TB/s versus 819 GB/s. More memory per GPU means larger models fit before you have to shard across cards, which often matters more than raw TFLOPS for inference.

For LLM training and inference, weight the FP8 and FP16 rows and memory capacity most heavily. For scientific and HPC workloads, the FP64 row is the one to read.