← Back to community benchmarks
Qwen3.5-122B-A10B9bit
Performance
1k
tokens
737.1
PP tok/s
43.9
TG tok/s
1389
TTFT (ms)
129.5
Peak mem (GB)
Hardware
Chip
M3 Ultra (80c)
Memory
256 GB
GPU Cores
80
Software
oMLX
v0.2.9
macOS
macOS 26.3.1
Context
1,024
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 737.1 | 43.9 | 129.5 GB | current |
| 4k | 868.2 | 42.4 | 129.8 GB | view |
| 8k | 872.8 | 38.7 | 130.1 GB | view |
| 16k | 832.7 | 30.3 | 130.8 GB | view |
| 32k | 741.7 | 21.9 | 132.2 GB | view |
| 64k | 576.6 | 14.2 | 134.9 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 43.9 | 1.00× |
| 2× | 68.0 | 1.55× |
| 4× | 96.7 | 2.20× |