← Back to community benchmarks
Qwen3.5-27B
Performance
16k
tokens
684.8
PP tok/s
15.9
TG tok/s
23924
TTFT (ms)
31.7
Peak mem (GB)
Hardware
Chip
M5 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.3.6
macOS
macOS 26.4.1
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 712.0 | 18.3 | 28.3 GB | view |
| 4k | 793.7 | 17.7 | 29.8 GB | view |
| 8k | 781.4 | 17.2 | 30.4 GB | view |
| 16k | 684.8 | 15.9 | 31.7 GB | current |
| 32k | 586.0 | 15.3 | 34.2 GB | view |
| 64k | 450.2 | 11.2 | 39.2 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 18.3 | 1.00× |
| 2× | 27.9 | 1.52× |
| 4× | 39.0 | 2.13× |