← Back to community benchmarks
Qwen3.6-27B-oQ8-mtp
Performance
8k
tokens
123.6
PP tok/s
6.8
TG tok/s
66293
TTFT (ms)
30.8
Peak mem (GB)
Hardware
Chip
M4 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.3.9
macOS
macOS 26.4
Context
8,192
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 102.6 | 7.1 | 28.3 GB | view |
| 4k | 124.0 | 6.9 | 29.8 GB | view |
| 8k | 123.6 | 6.8 | 30.8 GB | current |
| 16k | 119.9 | 6.6 | 32.3 GB | view |
| 32k | 115.2 | 14.8 | 35.3 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 7.1 | 1.00× |
| 2× | 31.7 | 4.46× |
| 4× | 22.7 | 3.20× |