← Back to community benchmarks
Qwen3.6-27B-oQ4-mtp
Performance
16k
tokens
311.5
PP tok/s
39.5
TG tok/s
52599
TTFT (ms)
20.7
Peak mem (GB)
Hardware
Chip
M3 Ultra (60c)
Memory
256 GB
GPU Cores
60
Software
oMLX
v0.3.9.dev1
macOS
macOS 26.5
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 301.8 | 43.2 | 16.7 GB | view |
| 4k | 321.2 | 44.4 | 18.2 GB | view |
| 8k | 319.8 | 42.9 | 19.2 GB | view |
| 16k | 311.5 | 39.5 | 20.7 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 43.2 | 1.00× |
| 2× | 57.6 | 1.33× |
| 4× | 65.2 | 1.51× |