← Back to community benchmarks
Qwen3.6-27B
Performance
8k
tokens
796.0
PP tok/s
17.5
TG tok/s
10291
TTFT (ms)
30.8
Peak mem (GB)
Hardware
Chip
M5 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.3.9rc1
macOS
macOS 26.4.1
Context
8,192
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 727.2 | 18.2 | 28.3 GB | view |
| 4k | 860.7 | 17.5 | 29.8 GB | view |
| 8k | 796.0 | 17.5 | 30.8 GB | current |
| 16k | 760.9 | 17.1 | 32.3 GB | view |
| 32k | 674.9 | 16.5 | 35.3 GB | view |
| 64k | 587.3 | 15.5 | 41.4 GB | view |
| 128k | 439.2 | 13.7 | 53.7 GB | view |
| 195k | 345.6 | 11.9 | 68.4 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 18.2 | 1.00× |
| 2× | 34.6 | 1.90× |
| 4× | 50.2 | 2.76× |
| 8× | 99.6 | 5.47× |