← Back to community benchmarks
Qwen3.5-35B-A3B
Performance
16k
tokens
2,151
PP tok/s
79.0
TG tok/s
7617
TTFT (ms)
22.0
Peak mem (GB)
Hardware
Chip
M5 Pro (20c)
Memory
64 GB
GPU Cores
20
Software
oMLX
v0.2.20.dev1
macOS
macOS 26.3.1
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 2,031 | 98.1 | 20.1 GB | view |
| 4k | 2,382 | 93.8 | 21.0 GB | view |
| 8k | 2,332 | 88.6 | 21.4 GB | view |
| 16k | 2,151 | 79.0 | 22.0 GB | current |
| 32k | 1,804 | 65.6 | 23.4 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 98.1 | 1.00× |
| 2× | 153.0 | 1.56× |
| 4× | 201.2 | 2.05× |