← Back to community benchmarks
Qwen3.6-27B-oQ6-fp16-mtp
Performance
16k
tokens
211.4
PP tok/s
17.9
TG tok/s
77504
TTFT (ms)
26.9
Peak mem (GB)
Hardware
Chip
M1 Ultra (48c)
Memory
64 GB
GPU Cores
48
Software
oMLX
v0.3.9
macOS
macOS 15.7.7
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 202.8 | 19.8 | 23.0 GB | view |
| 4k | 216.3 | 19.6 | 24.4 GB | view |
| 8k | 215.7 | 19.1 | 25.4 GB | view |
| 16k | 211.4 | 17.9 | 26.9 GB | current |
| 32k | 200.3 | 16.5 | 29.9 GB | view |
| 64k | 169.0 | 13.7 | 36.0 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 19.8 | 1.00× |
| 2× | 26.5 | 1.34× |
| 4× | 30.0 | 1.52× |
| 8× | 32.8 | 1.66× |