← Back to community benchmarks
Qwen3.6-27B-oQ4-mtp
Performance
16k
tokens
105.9
PP tok/s
17.1
TG tok/s
154680
TTFT (ms)
20.4
Peak mem (GB)
Hardware
Chip
M2 Max (38c)
Memory
64 GB
GPU Cores
38
Software
oMLX
v0.4.4
macOS
macOS 26.5
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 111.6 | 19.5 | 16.5 GB | view |
| 4k | 105.6 | 17.5 | 17.9 GB | view |
| 8k | 107.4 | 18.2 | 18.9 GB | view |
| 16k | 105.9 | 17.1 | 20.4 GB | current |
| 32k | 102.3 | 15.9 | 23.4 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 19.5 | 1.00× |
| 2× | 26.9 | 1.38× |
| 4× | 28.8 | 1.48× |
| 8× | 22.9 | 1.17× |