← Back to community benchmarks
Qwen3.6-27B-oQ8-mtp
Performance
16k
tokens
310.5
PP tok/s
29.6
TG tok/s
52775
TTFT (ms)
32.8
Peak mem (GB)
Hardware
Chip
M3 Ultra (60c)
Memory
256 GB
GPU Cores
60
Software
oMLX
v0.3.9
macOS
macOS 26.4.1
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 298.6 | 31.2 | 28.8 GB | view |
| 4k | 320.0 | 31.7 | 30.3 GB | view |
| 8k | 318.6 | 30.7 | 31.3 GB | view |
| 16k | 310.5 | 29.6 | 32.8 GB | current |
| 32k | 292.5 | 27.1 | 35.8 GB | view |
| 64k | 257.1 | 23.2 | 41.8 GB | view |
| 128k | 203.6 | 18.6 | 54.2 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 31.2 | 1.00× |
| 2× | 39.1 | 1.25× |
| 4× | 49.9 | 1.60× |