← Back to community benchmarks
Qwen3.6-27B-oQ6-mtp
Performance
195k
tokens
124.6
PP tok/s
13.0
TG tok/s
1604850
TTFT (ms)
63.0
Peak mem (GB)
Hardware
Chip
M4 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.3.9
macOS
macOS 15.7.7
Context
200,000
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 228.1 | 28.0 | 23.0 GB | view |
| 4k | 217.1 | 27.3 | 24.4 GB | view |
| 8k | 202.0 | 23.2 | 25.4 GB | view |
| 16k | 200.9 | 23.9 | 26.9 GB | view |
| 32k | 199.1 | 22.9 | 29.9 GB | view |
| 64k | 178.9 | 19.7 | 36.0 GB | view |
| 128k | 149.3 | 16.0 | 48.3 GB | view |
| 195k | 124.6 | 13.0 | 63.0 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 28.0 | 1.00× |
| 2× | 31.8 | 1.14× |
| 4× | 36.0 | 1.29× |
| 8× | 37.7 | 1.35× |