← Back to community benchmarks
Qwen3.6-27B-oQ4-mtp
Performance
4k
tokens
422.8
PP tok/s
32.6
TG tok/s
9688
TTFT (ms)
55.0
Peak mem (GB)
Hardware
Chip
M3 Ultra (80c)
Memory
512 GB
GPU Cores
80
Software
oMLX
v0.3.8
macOS
macOS 26.3.2
Context
4,096
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 393.1 | 31.8 | 53.6 GB | view |
| 4k | 422.8 | 32.6 | 55.0 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 31.8 | 1.00× |
| 2× | 33.3 | 1.05× |
| 4× | 44.5 | 1.40× |