← Back to community benchmarks
Qwen3.6-27B-UD
Performance
16k
tokens
1,223
PP tok/s
10.3
TG tok/s
13398
TTFT (ms)
29.4
Peak mem (GB)
Hardware
Chip
M3 Ultra (60c)
Memory
256 GB
GPU Cores
60
Software
oMLX
v0.3.8
macOS
macOS 26.4.1
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 300.9 | 26.0 | 25.6 GB | view |
| 4k | 320.7 | 25.5 | 27.0 GB | view |
| 8k | 319.0 | 24.9 | 28.1 GB | view |
| 16k | 1,223 | 10.3 | 29.4 GB | current |
| 32k | 1,203 | 10.2 | 30.7 GB | view |
| 64k | 1,120 | 10.0 | 32.5 GB | view |
| 128k | 955.2 | 9.7 | 36.3 GB | view |
| 195k | 806.3 | 9.3 | 40.6 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 26.0 | 1.00× |
| 2× | 26.5 | 1.02× |
| 4× | 5.0 | 0.19× |