← Back to community benchmarks
Qwen3.5-27B
Performance
16k
tokens
83.3
PP tok/s
12.8
TG tok/s
196687
TTFT (ms)
18.3
Peak mem (GB)
Hardware
Chip
M1 Max (32c)
Memory
64 GB
GPU Cores
32
Software
oMLX
v0.2.7
macOS
macOS 26.3.1
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 87.0 | 18.3 | 16.6 GB | view |
| 4k | 87.0 | 17.2 | 17.0 GB | view |
| 8k | 85.8 | 16.1 | 17.4 GB | view |
| 16k | 83.3 | 12.8 | 18.3 GB | current |
| 32k | 80.8 | 9.3 | 20.1 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 18.3 | 1.00× |
| 2× | 21.7 | 1.19× |
| 4× | 23.2 | 1.27× |
| 8× | 17.5 | 0.96× |