Performance
32k
tokens
3,138
PP tok/s
93.8
TG tok/s
10444
TTFT (ms)
23.6
Peak mem (GB)
Hardware
Chip
M5 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.5.1
macOS
macOS 26.5.2
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,614 | 110.6 | 19.9 GB | view |
| 4k | 3,539 | 108.0 | 20.7 GB | view |
| 8k | 4,351 | 104.0 | 21.1 GB | view |
| 16k | 3,834 | 101.2 | 22.0 GB | view |
| 32k | 3,138 | 93.8 | 23.6 GB | current |
| 64k | 2,231 | 81.7 | 25.9 GB | view |
| 128k | 1,306 | 64.0 | 31.5 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 110.6 | 1.00× |
| 2× | 173.7 | 1.57× |
| 4× | 271.1 | 2.45× |