How we got this number
Each generated token reads the whole model from memory once: 17 GB. The GEEKOM A6 (Ryzen 7 6800H, 32GB) moves 76.8 GB/s, so the ceiling is 4.5 tokens/sec. Real machines reach a fraction of that ceiling; we use 60%, an assumption until we get measured runs for this kind of machine. Send us your real numbers and we'll replace the estimate.
Faster machines for translateGemma 27B
| Mac Studio M5 Ultra (96GB) | 58 tok/s | Check price → |
| Mac Studio M5 Ultra (256GB) | 58 tok/s | Check price → |
| Mac Studio M5 Max (40-core GPU, 64GB) | 29 tok/s | Check price → |
| Mac mini M5 Pro (64GB) | 14 tok/s | Check price → |
Other models on the GEEKOM A6 (Ryzen 7 6800H, 32GB)
| lfm 2.5 8B (A1B) | Fast | 38 tok/s |
| Qwen3.5 4B | Comfortable | 14 tok/s |
| Qwen 3 30B (A3B) | Comfortable | 14 tok/s |
| Qwen3 Coder 30B (A3B) | Comfortable | 14 tok/s |
| Qwen 3-VL 30B (A3B) | Comfortable | 13 tok/s |
| Llama 2 7B | Comfortable | 12 tok/s |