How we got this number
Each generated token reads the active part of the model from memory once: 7.2 GB. The Mac mini M5 Pro (24GB) moves 307 GB/s, so the ceiling is 43 tokens/sec. Real machines reach a fraction of that ceiling; we use 80%, calibrated against a measured Mac run. Send us your real numbers and we'll replace the estimate.
Faster machines for mixtral 8X7B
| Mac Studio M5 Ultra (96GB) | 82 tok/s | Check price → |
| Mac Studio M5 Ultra (256GB) | 82 tok/s | Check price → |
| Mac Studio M5 Max (40-core GPU, 64GB) | 41 tok/s | Check price → |
| Mac mini M5 Pro (64GB) | 20 tok/s | Check price → |
Other models on the Mac mini M5 Pro (24GB)
| gpt-oss 20B | Fast | 60 tok/s |
| Granite 4.2 8B | Fast | 46 tok/s |
| Granite 4.1 8B | Fast | 46 tok/s |
| Gemma 2 9B | Fast | 46 tok/s |
| minicpm v 8B | Fast | 45 tok/s |
| LLaVA Llama3 8B | Fast | 45 tok/s |