← Back to community benchmarks
Qwen3.6-27B-MTPLX-Optimized-Speed
Performance
16k
tokens
90.0
PP tok/s
7.7
TG tok/s
182143
TTFT (ms)
19.8
Peak mem (GB)
Hardware
Chip
M3 Pro (18c)
Memory
36 GB
GPU Cores
18
Software
oMLX
v0.3.9.dev2
macOS
macOS 26.4.1
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 93.7 | 8.2 | 15.8 GB | view |
| 4k | 96.9 | 7.9 | 17.3 GB | view |
| 8k | 96.6 | 8.4 | 18.3 GB | view |
| 16k | 90.0 | 7.7 | 19.8 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 8.2 | 1.00× |
| 2× | 15.7 | 1.91× |
| 4× | 21.5 | 2.62× |
| 8× | 24.5 | 2.99× |