← Back to community benchmarks
Qwen3.5-122B-A10B9bit
Performance
1k
tokens
737.9
PP tok/s
41.1
TG tok/s
1388
TTFT (ms)
129.6
Peak mem (GB)
Hardware
Chip
M3 Ultra (80c)
Memory
512 GB
GPU Cores
80
Software
oMLX
v0.2.19
macOS
macOS 26.3.1
Context
1,024
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 737.9 | 41.1 | 129.6 GB | current |
| 8k | 912.7 | 38.1 | 131.6 GB | view |
| 16k | 867.3 | 35.7 | 132.8 GB | view |
| 32k | 746.2 | 31.0 | 135.2 GB | view |
| 64k | 576.2 | 26.3 | 140.1 GB | view |
| 128k | 376.8 | 18.8 | 149.7 GB | view |
| 195k | 275.2 | 14.2 | 159.9 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 41.1 | 1.00× |
| 2× | 64.3 | 1.56× |
| 4× | 89.4 | 2.18× |