← Back to community benchmarks
Qwen3.6-35B-A3B
Performance
32k
tokens
1,772
PP tok/s
73.6
TG tok/s
18492
TTFT (ms)
23.1
Peak mem (GB)
Hardware
Chip
M5 Pro (20c)
Memory
64 GB
GPU Cores
20
Software
oMLX
v0.5.1
macOS
macOS 26.5.2
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 2,084 | 95.9 | 19.3 GB | view |
| 4k | 2,224 | 91.8 | 20.1 GB | view |
| 8k | 2,227 | 89.0 | 20.5 GB | view |
| 16k | 2,071 | 82.9 | 21.4 GB | view |
| 32k | 1,772 | 73.6 | 23.1 GB | current |
| 64k | 1,325 | 59.6 | 26.0 GB | view |
| 128k | 783.1 | 41.2 | 31.7 GB | view |
| 195k | 532.7 | 32.2 | 39.4 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 95.9 | 1.00× |
| 2× | 153.4 | 1.60× |
| 4× | 214.4 | 2.24× |
| 8× | 402.0 | 4.19× |