← Back to community benchmarks
Qwen3.6-35B-A3B
Performance
64k
tokens
1,336
PP tok/s
42.4
TG tok/s
49068
TTFT (ms)
25.0
Peak mem (GB)
Hardware
Chip
M5 Pro (20c)
Memory
64 GB
GPU Cores
20
Software
oMLX
v0.3.6
macOS
macOS 26.4.1
Context
65,536
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,497 | 99.2 | 19.3 GB | view |
| 4k | 2,212 | 93.3 | 20.0 GB | view |
| 8k | 2,259 | 88.8 | 20.4 GB | view |
| 16k | 2,143 | 80.8 | 21.0 GB | view |
| 32k | 1,828 | 67.9 | 22.4 GB | view |
| 64k | 1,336 | 42.4 | 25.0 GB | current |
| 128k | 839.8 | 27.0 | 30.4 GB | view |
| 195k | 608.2 | 19.3 | 36.1 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 99.2 | 1.00× |
| 2× | 137.9 | 1.39× |
| 4× | 179.8 | 1.81× |
| 8× | 220.9 | 2.23× |