← Back to community benchmarks
Qwen3.5-35B-A3B
Performance
8k
tokens
727.5
PP tok/s
51.6
TG tok/s
11260
TTFT (ms)
21.4
Peak mem (GB)
Hardware
Chip
M4 Pro (16c)
Memory
48 GB
GPU Cores
16
Software
oMLX
v0.2.21
macOS
macOS 26.3.1
Context
8,192
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 730.2 | 59.8 | 20.1 GB | view |
| 4k | 761.0 | 56.1 | 21.0 GB | view |
| 8k | 727.5 | 51.6 | 21.4 GB | current |
| 16k | 669.9 | 43.4 | 22.0 GB | view |
| 32k | 562.8 | 34.4 | 23.4 GB | view |
| 64k | 447.1 | 21.2 | 26.0 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 59.8 | 1.00× |
| 2× | 92.0 | 1.54× |
| 4× | 113.9 | 1.90× |