← Back to community benchmarks
Qwen3.6-27B
Performance
32k
tokens
477.4
PP tok/s
16.8
TG tok/s
68636
TTFT (ms)
31.3
Peak mem (GB)
Hardware
Chip
M5 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.5.3
macOS
macOS 26.5.2
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 826.7 | 18.3 | 28.3 GB | view |
| 4k | 884.0 | 18.0 | 29.8 GB | view |
| 8k | 827.0 | 17.8 | 30.4 GB | view |
| 16k | 623.7 | 17.3 | 30.3 GB | view |
| 32k | 477.4 | 16.8 | 31.3 GB | current |
| 64k | 357.8 | 15.8 | 33.4 GB | view |
| 128k | 234.7 | 13.8 | 37.7 GB | view |
| 195k | 175.1 | 12.4 | 42.1 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 18.3 | 1.00× |
| 2× | 37.0 | 2.02× |
| 4× | 73.9 | 4.04× |
| 8× | 147.7 | 8.07× |