← Back to community benchmarks
Qwen3.6-27B-oQ4-mtp
Performance
32k
tokens
171.3
PP tok/s
25.5
TG tok/s
191270
TTFT (ms)
23.7
Peak mem (GB)
Hardware
Chip
M2 Ultra (60c)
Memory
128 GB
GPU Cores
60
Software
oMLX
v0.3.9
macOS
macOS 26.4.1
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 175.5 | 31.9 | 16.8 GB | view |
| 4k | 182.3 | 31.8 | 18.2 GB | view |
| 8k | 181.7 | 29.8 | 19.2 GB | view |
| 16k | 178.5 | 27.9 | 20.7 GB | view |
| 32k | 171.3 | 25.5 | 23.7 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 31.9 | 1.00× |
| 2× | 39.0 | 1.22× |
| 4× | 43.4 | 1.36× |
| 8× | 46.6 | 1.46× |