← Back to community benchmarks
Qwen3.5-35B-A3B
Performance
32k
tokens
1,588
PP tok/s
48.3
TG tok/s
20641
TTFT (ms)
37.8
Peak mem (GB)
Hardware
Chip
M5 Pro (20c)
Memory
64 GB
GPU Cores
20
Software
oMLX
v0.2.11
macOS
macOS 26.3.1
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,741 | 67.0 | 36.2 GB | view |
| 4k | 1,846 | 64.7 | 36.3 GB | view |
| 8k | 1,945 | 61.0 | 36.5 GB | view |
| 16k | 1,796 | 56.4 | 37.0 GB | view |
| 32k | 1,588 | 48.3 | 37.8 GB | current |
| 64k | 1,192 | 36.7 | 39.5 GB | view |
| 128k | 786.2 | 19.9 | 42.8 GB | view |
| 195k | 574.1 | 8.4 | 46.4 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 67.0 | 1.00× |
| 2× | 104.6 | 1.56× |
| 4× | 144.5 | 2.16× |
| 8× | 186.3 | 2.78× |