Performance
8k
tokens
163.1
PP tok/s
14.8
TG tok/s
50228
TTFT (ms)
398.0
Peak mem (GB)
Hardware
Chip
M3 Ultra (80c)
Memory
512 GB
GPU Cores
80
Software
oMLX
v0.2.11a1
macOS
macOS 15.5
Context
8,192
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 8k | 163.1 | 14.8 | 398.0 GB | current |
| 16k | 127.2 | 13.8 | 404.0 GB | view |