← Back to community benchmarks
Qwen3.5-35B-A3B
Performance
32k
tokens
662.9
PP tok/s
50.6
TG tok/s
49428
TTFT (ms)
21.6
Peak mem (GB)
Hardware
Chip
M4 Pro (20c)
Memory
64 GB
GPU Cores
20
Software
oMLX
v0.2.6
macOS
macOS 26.0.1
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 826.7 | 86.3 | 20.1 GB | view |
| 4k | 894.9 | 80.8 | 20.2 GB | view |
| 8k | 875.1 | 75.2 | 20.4 GB | view |
| 16k | 806.3 | 64.1 | 20.8 GB | view |
| 32k | 662.9 | 50.6 | 21.6 GB | current |
| 64k | 511.5 | 33.2 | 23.3 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 86.3 | 1.00× |
| 2× | 128.2 | 1.49× |
| 4× | 162.8 | 1.89× |