← Back to community benchmarks
Qwen3.6-27B
Performance
32k
tokens
360.8
PP tok/s
10.2
TG tok/s
90824
TTFT (ms)
27.9
Peak mem (GB)
Hardware
Chip
M5 Pro (20c)
Memory
48 GB
GPU Cores
20
Software
oMLX
v0.3.6
macOS
macOS 26.4.1
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 395.6 | 12.8 | 22.1 GB | view |
| 4k | 441.6 | 12.6 | 23.5 GB | view |
| 8k | 425.3 | 12.2 | 24.2 GB | view |
| 16k | 374.3 | 11.4 | 25.4 GB | view |
| 32k | 360.8 | 10.2 | 27.9 GB | current |
| 64k | 286.9 | 8.7 | 32.9 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 12.8 | 1.00× |
| 2× | 19.0 | 1.48× |
| 4× | 26.1 | 2.04× |