← Back to community benchmarks
Qwen3.6-27B
Performance
8k
tokens
387.3
PP tok/s
17.7
TG tok/s
21154
TTFT (ms)
18.3
Peak mem (GB)
Hardware
Chip
M5 Pro (16c)
Memory
48 GB
GPU Cores
16
Software
oMLX
v0.4.3
macOS
macOS 26.5.1
Context
8,192
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 333.5 | 25.1 | 15.8 GB | view |
| 4k | 377.2 | 17.3 | 16.7 GB | view |
| 8k | 387.3 | 17.7 | 18.3 GB | current |
| 16k | 359.8 | 16.8 | 19.8 GB | view |
| 32k | 325.5 | 15.4 | 22.8 GB | view |
| 64k | 158.1 | 12.9 | 26.7 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 25.1 | 1.00× |
| 2× | 26.9 | 1.07× |
| 4× | 32.0 | 1.27× |