← Back to community benchmarks
Qwen3.6-35B-A3B
Performance
32k
tokens
2,215
PP tok/s
79.7
TG tok/s
14794
TTFT (ms)
29.6
Peak mem (GB)
Hardware
Chip
M3 Ultra (80c)
Memory
512 GB
GPU Cores
80
Software
oMLX
v0.3.8
macOS
macOS 26.4.1
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,679 | 103.9 | 26.5 GB | view |
| 4k | 2,555 | 101.7 | 27.3 GB | view |
| 8k | 2,713 | 97.1 | 27.7 GB | view |
| 16k | 2,570 | 88.1 | 28.5 GB | view |
| 32k | 2,215 | 79.7 | 29.6 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 103.9 | 1.00× |
| 2× | 203.2 | 1.96× |
| 4× | 403.5 | 3.88× |
| 8× | 833.3 | 8.02× |