← Back to community benchmarks
Qwen3.5-122B-A10B
Performance
16k
tokens
655.3
PP tok/s
29.8
TG tok/s
25003
TTFT (ms)
124.5
Peak mem (GB)
Hardware
Chip
M3 Ultra (60c)
Memory
256 GB
GPU Cores
60
Software
oMLX
v0.2.12
macOS
macOS 26.3.1
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 671.1 | 41.2 | 123.2 GB | view |
| 4k | 704.5 | 39.7 | 123.5 GB | view |
| 8k | 695.4 | 37.9 | 123.9 GB | view |
| 16k | 655.3 | 29.8 | 124.5 GB | current |
| 32k | 579.4 | 22.2 | 125.9 GB | view |
| 64k | 460.6 | 15.2 | 128.7 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 41.2 | 1.00× |
| 2× | 66.0 | 1.60× |
| 4× | 91.3 | 2.22× |