← Back to community benchmarks
Qwen3.6-35B-A3B
Performance
32k
tokens
667.5
PP tok/s
54.3
TG tok/s
49093
TTFT (ms)
22.7
Peak mem (GB)
Hardware
Chip
M4 Pro (20c)
Memory
64 GB
GPU Cores
20
Software
oMLX
v0.3.8
macOS
macOS 15.3.1
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 758.6 | 83.7 | 19.3 GB | view |
| 4k | 888.4 | 80.0 | 20.0 GB | view |
| 8k | 871.4 | 74.8 | 20.5 GB | view |
| 16k | 810.3 | 66.3 | 21.2 GB | view |
| 32k | 667.5 | 54.3 | 22.7 GB | current |
| 64k | 513.3 | 33.2 | 25.0 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 83.7 | 1.00× |
| 2× | 127.8 | 1.53× |
| 4× | 159.6 | 1.91× |
| 8× | 194.5 | 2.32× |