Home › GPUs & Macs › Apple M1 Ultra (64-core GPU) vs Apple M2 Ultra (60-core GPU)
Apple M1 Ultra (64-core GPU) vs Apple M2 Ultra (60-core GPU) for local LLMs
Across 6 tests measured on both, the Apple M2 Ultra (60-core GPU) is faster in 6 and generates a median 1.07× the tokens/s of the Apple M1 Ultra (64-core GPU).
On Llama 2 7B (F16, llama-bench) the Apple M1 Ultra (64-core GPU) generates 37.0 tokens/s and the Apple M2 Ultra (60-core GPU) 39.9 tokens/s (measured).
Same model, same quant, same test
| Model | Quant / test | M1 Ultra (64-core GPU) tok/s | M2 Ultra (60-core GPU) tok/s | Faster | Sources and run details |
|---|---|---|---|---|---|
| Llama 2 7B | F16 llama-bench | 37.0 Metal | 39.9 Metal | M2 Ultra (60-core GPU) +8% | llama.cpp #4167 M1 Ultra (64-core GPU) · llama.cpp #4167 M2 Ultra (60-core GPU) |
| Llama 2 7B | Q4_0 llama-bench | 83.7 Metal | 88.6 Metal | M2 Ultra (60-core GPU) +6% | llama.cpp #4167 M1 Ultra (64-core GPU) · llama.cpp #4167 M2 Ultra (60-core GPU) |
| Llama 2 7B | Q8_0 llama-bench | 59.9 Metal | 62.1 Metal | M2 Ultra (60-core GPU) +4% | llama.cpp #4167 M1 Ultra (64-core GPU) · llama.cpp #4167 M2 Ultra (60-core GPU) |
| Llama 3.1 8B Instruct | Q4_K_M LocalScore | 54.3 llamafile | 59.4 llamafile | M2 Ultra (60-core GPU) +9% | LocalScore M1 Ultra (64-core GPU) · LocalScore M2 Ultra (60-core GPU) |
| Llama 3.2 1B Instruct | Q4_K_M LocalScore | 151 llamafile | 174 llamafile | M2 Ultra (60-core GPU) +15% | LocalScore M1 Ultra (64-core GPU) · LocalScore M2 Ultra (60-core GPU) |
| Qwen2.5 14B Instruct | Q4_K_M LocalScore | 32.4 llamafile | 34.1 llamafile | M2 Ultra (60-core GPU) +5% | LocalScore M1 Ultra (64-core GPU) · LocalScore M2 Ultra (60-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 M1 Ultra (64-core GPU) | Apple M2 Ultra (60-core GPU) | |
|---|---|---|
| Memory | Unified memory | Unified memory |
| Memory bandwidth | 800 GB/s | 800 GB/s |
All results for the Apple M1 Ultra (64-core GPU) · All results for the Apple M2 Ultra (60-core GPU) · More comparisons