← Back to community benchmarks
Qwen3.6-27B-oQ4-mtp
Performance
64k
tokens
90.8
PP tok/s
13.7
TG tok/s
721371
TTFT (ms)
27.6
Peak mem (GB)
Hardware
Chip
M4 Pro (20c)
Memory
64 GB
GPU Cores
20
Software
oMLX
v0.3.9.dev2
macOS
macOS 26.4.1
Context
65,536
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 122.3 | 22.2 | 16.8 GB | view |
| 4k | 105.1 | 18.9 | 18.2 GB | view |
| 64k | 90.8 | 13.7 | 27.6 GB | current |
| 128k | 75.3 | 9.8 | 42.1 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 22.2 | 1.00× |
| 2× | 25.4 | 1.14× |
| 4× | 27.5 | 1.24× |