← Back to community benchmarks
Qwen3.6-27B-oQ8-mtp
Performance
64k
tokens
442.9
PP tok/s
17.7
TG tok/s
147964
TTFT (ms)
41.8
Peak mem (GB)
Hardware
Chip
M5 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.4.1
macOS
macOS 26.5
Context
65,536
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 705.0 | 29.2 | 28.8 GB | view |
| 4k | 818.6 | 27.6 | 30.3 GB | view |
| 8k | 774.0 | 26.4 | 31.3 GB | view |
| 16k | 679.9 | 25.7 | 32.8 GB | view |
| 32k | 613.0 | 22.1 | 35.8 GB | view |
| 64k | 442.9 | 17.7 | 41.8 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 29.2 | 1.00× |
| 2× | 26.2 | 0.90× |
| 4× | 36.7 | 1.26× |