Home › GPUs & Macs › Apple M4 Pro (16-core GPU) vs Apple M5 Pro (20-core GPU)

Apple M4 Pro (16-core GPU) vs Apple M5 Pro (20-core GPU) for local LLMs

Across 6 tests measured on both, the Apple M5 Pro (20-core GPU) is faster in 6 and generates a median 1.41× the tokens/s of the Apple M4 Pro (16-core GPU).

On Llama 2 7B (F16, llama-bench) the Apple M4 Pro (16-core GPU) generates 17.2 tokens/s and the Apple M5 Pro (20-core GPU) 21.6 tokens/s (measured).

Same model, same quant, same test

ModelQuant / testM4 Pro (16-core GPU)
tok/s
M5 Pro (20-core GPU)
tok/s
FasterSources and run details
Llama 2 7BF16
llama-bench
17.2
Metal
21.6
Metal
M5 Pro (20-core GPU) +25%llama.cpp #4167 M4 Pro (16-core GPU) · llama.cpp #4167 M5 Pro (20-core GPU)
Llama 2 7BQ4_0
llama-bench
49.6
Metal
66.3
Metal
M5 Pro (20-core GPU) +34%llama.cpp #4167 M4 Pro (16-core GPU) · llama.cpp #4167 M5 Pro (20-core GPU)
Llama 2 7BQ8_0
llama-bench
30.5
Metal
38.9
Metal
M5 Pro (20-core GPU) +27%llama.cpp #4167 M4 Pro (16-core GPU) · llama.cpp #4167 M5 Pro (20-core GPU)
Llama 3.1 8B InstructQ4_K_M
LocalScore
30.2
llamafile
44.7
llamafile
M5 Pro (20-core GPU) +48%LocalScore M4 Pro (16-core GPU) · LocalScore M5 Pro (20-core GPU)
Llama 3.2 1B InstructQ4_K_M
LocalScore
111
llamafile
175
llamafile
M5 Pro (20-core GPU) +58%LocalScore M4 Pro (16-core GPU) · LocalScore M5 Pro (20-core GPU)
Qwen2.5 14B InstructQ4_K_M
LocalScore
16.1
llamafile
24.0
llamafile
M5 Pro (20-core GPU) +49%LocalScore M4 Pro (16-core GPU) · LocalScore M5 Pro (20-core GPU)

Generation speed in tokens/s, best measured backend for each chip (shown under the number). Backends can differ between the two chips; check them before drawing conclusions. Measured results only, nothing estimated.

Specs

Apple M4 Pro (16-core GPU)Apple M5 Pro (20-core GPU)
MemoryUnified memoryUnified memory
Memory bandwidth273 GB/s307 GB/s

All results for the Apple M4 Pro (16-core GPU) · All results for the Apple M5 Pro (20-core GPU) · More comparisons