← Back to community benchmarks
Qwen3.5-122B-A10B
Performance
4k
tokens
1,429
PP tok/s
60.8
TG tok/s
2867
TTFT (ms)
66.7
Peak mem (GB)
Hardware
Chip
M5 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.2.12
macOS
macOS 26.3.1
Context
4,096
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,261 | 60.7 | 66.3 GB | view |
| 4k | 1,429 | 60.8 | 66.7 GB | current |
| 8k | 1,391 | 57.2 | 67.0 GB | view |
| 16k | 1,277 | 51.2 | 67.7 GB | view |
| 32k | 1,004 | 36.8 | 69.1 GB | view |
| 64k | 789.7 | 23.1 | 71.9 GB | view |
| 128k | 545.4 | 12.5 | 77.4 GB | view |
| 195k | 410.4 | 8.9 | 83.4 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 60.7 | 1.00× |
| 2× | 98.2 | 1.62× |
| 4× | 134.4 | 2.21× |
| 8× | 167.5 | 2.76× |