Home › GPUs & Macs › Apple M4 (10-core GPU) vs Apple M4 (8-core GPU)
Apple M4 (10-core GPU) vs Apple M4 (8-core GPU) for local LLMs
Across 3 tests measured on both, the Apple M4 (10-core GPU) is faster in 3 and generates a median 1.13× the tokens/s of the Apple M4 (8-core GPU).
On Llama 3.1 8B Instruct (Q4_K_M, LocalScore) the Apple M4 (10-core GPU) generates 16.8 tokens/s and the Apple M4 (8-core GPU) 15.3 tokens/s (measured).
Same model, same quant, same test
| Model | Quant / test | M4 (10-core GPU) tok/s | M4 (8-core GPU) tok/s | Faster | Sources and run details |
|---|---|---|---|---|---|
| Llama 3.1 8B Instruct | Q4_K_M LocalScore | 16.8 llamafile | 15.3 llamafile | M4 (10-core GPU) +10% | LocalScore M4 (10-core GPU) · LocalScore M4 (8-core GPU) |
| Llama 3.2 1B Instruct | Q4_K_M LocalScore | 74.3 llamafile | 65.9 llamafile | M4 (10-core GPU) +13% | LocalScore M4 (10-core GPU) · LocalScore M4 (8-core GPU) |
| Qwen2.5 14B Instruct | Q4_K_M LocalScore | 9.0 llamafile | 7.3 llamafile | M4 (10-core GPU) +24% | LocalScore M4 (10-core GPU) · LocalScore M4 (8-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 (10-core GPU) | Apple M4 (8-core GPU) | |
|---|---|---|
| Memory | Unified memory | Unified memory |
| Memory bandwidth | 120 GB/s | 120 GB/s |
All results for the Apple M4 (10-core GPU) · All results for the Apple M4 (8-core GPU) · More comparisons