← Back to community benchmarks
Qwen3.6-27B-oQ4-mtp
Performance
32k
tokens
276.5
PP tok/s
33.1
TG tok/s
118512
TTFT (ms)
23.7
Peak mem (GB)
Hardware
Chip
M3 Ultra (60c)
Memory
256 GB
GPU Cores
60
Software
oMLX
v0.3.9.dev1
macOS
macOS 26.3.1
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 300.7 | 44.6 | 16.8 GB | view |
| 4k | 320.0 | 43.1 | 18.2 GB | view |
| 8k | 319.1 | 40.8 | 19.2 GB | view |
| 16k | 301.1 | 37.7 | 20.7 GB | view |
| 32k | 276.5 | 33.1 | 23.7 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 44.6 | 1.00× |
| 2× | 53.5 | 1.20× |
| 4× | 64.6 | 1.45× |