← Back to community benchmarks
Qwen3.5-122B-A10B
Performance
16k
tokens
826.4
PP tok/s
32.5
TG tok/s
19826
TTFT (ms)
124.5
Peak mem (GB)
Hardware
Chip
M3 Ultra (80c)
Memory
512 GB
GPU Cores
80
Software
oMLX
v0.2.6
macOS
macOS 15.5
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 837.2 | 39.1 | 123.2 GB | view |
| 4k | 895.1 | 37.8 | 123.5 GB | view |
| 8k | 876.6 | 35.7 | 123.9 GB | view |
| 16k | 826.4 | 32.5 | 124.5 GB | current |
| 32k | 717.8 | 27.9 | 125.9 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 39.1 | 1.00× |
| 2× | 62.4 | 1.60× |
| 4× | 89.0 | 2.28× |
| 8× | 129.4 | 3.31× |