← Back to community benchmarks
Qwen3.6-35B-A3B
Performance
8k
tokens
1,695
PP tok/s
106.7
TG tok/s
4832
TTFT (ms)
20.5
Peak mem (GB)
Hardware
Chip
M4 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.3.8
macOS
macOS 26.4.1
Context
8,192
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,377 | 112.1 | 19.3 GB | view |
| 4k | 1,705 | 110.7 | 20.0 GB | view |
| 8k | 1,695 | 106.7 | 20.5 GB | current |
| 16k | 1,587 | 97.9 | 21.0 GB | view |
| 32k | 1,376 | 85.8 | 22.7 GB | view |
| 64k | 1,059 | 69.9 | 25.0 GB | view |
| 128k | 692.8 | 47.9 | 30.4 GB | view |
| 195k | 486.2 | 12.6 | 38.6 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 112.1 | 1.00× |
| 2× | 181.6 | 1.62× |
| 4× | 274.1 | 2.45× |