← Back to community benchmarks
Qwen3.6-27B-oQ4-mtp
Performance
16k
tokens
395.3
PP tok/s
22.6
TG tok/s
41450
TTFT (ms)
20.0
Peak mem (GB)
Hardware
Chip
M5 Pro (20c)
Memory
64 GB
GPU Cores
20
Software
oMLX
v0.3.9
macOS
macOS 26.5
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 418.4 | 26.1 | 16.7 GB | view |
| 4k | 449.4 | 25.3 | 18.2 GB | view |
| 8k | 428.0 | 24.5 | 18.8 GB | view |
| 16k | 395.3 | 22.6 | 20.0 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 26.1 | 1.00× |
| 2× | 29.2 | 1.12× |
| 4× | 36.1 | 1.38× |