← Back to community benchmarks
Qwen3.5-27B
Performance
32k
tokens
112.9
PP tok/s
9.2
TG tok/s
290120
TTFT (ms)
21.8
Peak mem (GB)
Hardware
Chip
M3 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.2.23
macOS
macOS 26.3.1
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 210.4 | 23.0 | 15.8 GB | view |
| 4k | 216.1 | 21.8 | 17.4 GB | view |
| 8k | 189.7 | 21.3 | 18.0 GB | view |
| 16k | 170.3 | 8.6 | 19.3 GB | view |
| 32k | 112.9 | 9.2 | 21.8 GB | current |
| 64k | 105.9 | 6.8 | 26.9 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 23.0 | 1.00× |
| 2× | 16.1 | 0.70× |
| 4× | 22.8 | 0.99× |