← Back to community benchmarks
Qwen3.6-35B-A3BMTP
Performance
8k
tokens
2,272
PP tok/s
93.1
TG tok/s
3606
TTFT (ms)
21.0
Peak mem (GB)
Hardware
Chip
M5 Pro (20c)
Memory
48 GB
GPU Cores
20
Software
oMLX
v0.3.9.dev2
macOS
macOS 26.4.1
Context
8,192
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,487 | 104.2 | 19.7 GB | view |
| 4k | 2,206 | 99.5 | 20.5 GB | view |
| 8k | 2,272 | 93.1 | 21.0 GB | current |
| 16k | 2,144 | 85.0 | 21.7 GB | view |
| 32k | 1,837 | 76.7 | 23.2 GB | view |
| 64k | 1,349 | 59.3 | 26.2 GB | view |
| 128k | 858.4 | 42.9 | 32.2 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 104.2 | 1.00× |
| 2× | 203.5 | 1.95× |
| 4× | 472.7 | 4.54× |