← Back to community benchmarks
Qwen3.6-27B-oQ6-mtp
Performance
16k
tokens
305.4
PP tok/s
29.0
TG tok/s
53642
TTFT (ms)
26.9
Peak mem (GB)
Hardware
Chip
M3 Ultra (60c)
Memory
256 GB
GPU Cores
60
Software
oMLX
v0.3.9.dev2
macOS
macOS 26.3
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 297.6 | 31.5 | 23.0 GB | view |
| 4k | 315.0 | 31.6 | 24.4 GB | view |
| 8k | 313.5 | 31.3 | 25.4 GB | view |
| 16k | 305.4 | 29.0 | 26.9 GB | current |
| 32k | 288.0 | 26.9 | 29.9 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 31.5 | 1.00× |
| 2× | 38.9 | 1.23× |
| 4× | 43.4 | 1.38× |
| 8× | 47.8 | 1.52× |