← Back to community benchmarks
Qwen3.6-27B-oQ4-mtp
Performance
64k
tokens
354.3
PP tok/s
16.9
TG tok/s
184985
TTFT (ms)
29.5
Peak mem (GB)
Hardware
Chip
M5 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.3.9.dev1
macOS
macOS 26.4.1
Context
65,536
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 727.6 | 31.0 | 16.5 GB | view |
| 4k | 665.1 | 30.4 | 17.9 GB | view |
| 8k | 613.0 | 27.2 | 18.9 GB | view |
| 16k | 553.2 | 25.3 | 20.4 GB | view |
| 32k | 519.0 | 22.6 | 23.4 GB | view |
| 64k | 354.3 | 16.9 | 29.5 GB | current |
| 128k | 268.6 | 13.2 | 41.8 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 31.0 | 1.00× |
| 2× | 32.7 | 1.05× |
| 4× | 40.3 | 1.30× |