How we got this number
Each generated token reads the whole model from memory once: 28 GB. The Mac mini M4 (16GB) moves 120 GB/s, so the ceiling is 4.3 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 Nemotron 3 33B
| Mac Studio M5 Ultra (96GB) | 35 tok/s | Check price → |
| Mac Studio M5 Ultra (256GB) | 35 tok/s | Check price → |
| Mac Studio M5 Max (40-core GPU, 64GB) | 18 tok/s | Check price → |
| Mac mini M5 Pro (64GB) | 8.8 tok/s | Check price → |
Other models on the Mac mini M4 (16GB)
| ornith 9B | Comfortable | 17 tok/s |
| Gemma 3n E2B | Comfortable | 17 tok/s |
| Qwen2.5-VL 7B | Comfortable | 16 tok/s |
| minicpm v4.5 8B | Comfortable | 16 tok/s |
| Qwen 3-VL 8B | Comfortable | 16 tok/s |
| falcon 3 10B | Comfortable | 15 tok/s |