← Back to community benchmarks
Qwen3.6-27B4bit-MTP
Performance
8k
tokens
428.0
PP tok/s
25.8
TG tok/s
19140
TTFT (ms)
18.5
Peak mem (GB)
Hardware
Chip
M5 Pro (20c)
Memory
64 GB
GPU Cores
20
Software
oMLX
v0.3.9.dev2
macOS
macOS 26.4.1
Context
8,192
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 426.0 | 25.9 | 16.1 GB | view |
| 4k | 468.7 | 27.0 | 17.5 GB | view |
| 8k | 428.0 | 25.8 | 18.5 GB | current |
| 16k | 392.9 | 23.8 | 20.0 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 25.9 | 1.00× |
| 2× | 32.1 | 1.24× |
| 4× | 42.3 | 1.63× |
| 8× | 45.7 | 1.76× |