← Back to community benchmarks
Qwen3.6-27B-oQ8-mtp
Performance
32k
tokens
645.4
PP tok/s
26.1
TG tok/s
50770
TTFT (ms)
35.8
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 | 699.5 | 29.9 | 28.8 GB | view |
| 4k | 837.4 | 28.5 | 30.3 GB | view |
| 8k | 806.3 | 28.1 | 31.3 GB | view |
| 16k | 726.7 | 26.7 | 32.8 GB | view |
| 32k | 645.4 | 26.1 | 35.8 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 29.9 | 1.00× |
| 2× | 35.3 | 1.18× |
| 4× | 47.9 | 1.60× |