How we got this number
Each generated token reads the whole model from memory once: 23 GB. The Mac mini M4 (16GB) moves 120 GB/s, so the ceiling is 5.2 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 ornith 1.5 35B
| Mac Studio M5 Ultra (96GB) | 43 tok/s | Check price → |
| Mac Studio M5 Ultra (256GB) | 43 tok/s | Check price → |
| Mac Studio M5 Max (40-core GPU, 64GB) | 21 tok/s | Check price → |
| Mac mini M5 Pro (64GB) | 11 tok/s | Check price → |
Other models on the Mac mini M4 (16GB)
| lfm 2.5 8B (A1B) | Fast | 78 tok/s |
| falcon 3 7B | Fast | 21 tok/s |
| DeepSeek-R1 7B | Fast | 20 tok/s |
| Qwen2.5 7B | Fast | 20 tok/s |
| Llama 3 8B | Fast | 20 tok/s |
| Qwen2.5 Coder 7B | Fast | 20 tok/s |