← Back to community benchmarks
Qwen3.6-27B-oQ4-mtp
Performance
1k
tokens
170.5
PP tok/s
28.7
TG tok/s
6007
TTFT (ms)
16.7
Peak mem (GB)
Hardware
Chip
M2 Ultra (60c)
Memory
64 GB
GPU Cores
60
Software
oMLX
v0.3.9rc1
macOS
macOS 26.3.1
Context
1,024
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 170.5 | 28.7 | 16.7 GB | current |
| 4k | 178.9 | 29.6 | 18.2 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 28.7 | 1.00× |
| 2× | 37.4 | 1.30× |