← Back to community benchmarks
Qwen3.6-27B-Q8mtp
Performance
1k
tokens
386.2
PP tok/s
32.3
TG tok/s
2651
TTFT (ms)
28.8
Peak mem (GB)
Hardware
Chip
M3 Ultra (80c)
Memory
512 GB
GPU Cores
80
Software
oMLX
v0.3.9
macOS
macOS 26.5
Context
1,024
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 386.2 | 32.3 | 28.8 GB | current |
| 4k | 420.5 | 31.9 | 30.3 GB | view |
| 8k | 420.1 | 30.4 | 31.3 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 32.3 | 1.00× |
| 2× | 38.3 | 1.19× |
| 4× | 57.2 | 1.77× |
| 8× | 151.7 | 4.70× |