← Back to community benchmarks
Qwen3.6-27B-oQ4-fp16-mtp
Performance
32k
tokens
666.2
PP tok/s
20.6
TG tok/s
49189
TTFT (ms)
22.3
Peak mem (GB)
Hardware
Chip
M5 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.3.8
macOS
macOS 26.5
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 749.5 | 22.3 | 16.4 GB | view |
| 4k | 876.2 | 29.2 | 17.9 GB | view |
| 8k | 822.5 | 26.7 | 18.5 GB | view |
| 16k | 721.9 | 27.8 | 20.4 GB | view |
| 32k | 666.2 | 20.6 | 22.3 GB | current |
| 64k | 526.0 | 15.3 | 29.4 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 22.3 | 1.00× |
| 2× | 37.8 | 1.70× |
| 4× | 44.3 | 1.99× |