← Back to community benchmarks
Qwen3.6-27B-oQ4-mtp
Performance
32k
tokens
646.4
PP tok/s
15.2
TG tok/s
50693
TTFT (ms)
46.4
Peak mem (GB)
Hardware
Chip
M5 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.3.9.dev2
macOS
macOS 26.4.1
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 765.8 | 46.4 | 16.8 GB | view |
| 4k | 876.1 | 44.6 | 18.2 GB | view |
| 8k | 833.8 | 44.4 | 19.2 GB | view |
| 16k | 735.0 | 40.1 | 20.7 GB | view |
| 32k | 646.4 | 15.2 | 46.4 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 46.4 | 1.00× |
| 2× | 35.5 | 0.77× |
| 4× | 50.0 | 1.08× |