← Back to community benchmarks
Qwen3.5-122B-A10B-oQ4
Performance
8k
tokens
636.7
PP tok/s
15.8
TG tok/s
12866
TTFT (ms)
73.3
Peak mem (GB)
Hardware
Chip
M3 Ultra (60c)
Memory
256 GB
GPU Cores
60
Software
oMLX
v0.2.20
macOS
macOS 26.3.1
Context
8,192
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 585.1 | 17.4 | 69.2 GB | view |
| 4k | 647.1 | 16.3 | 72.1 GB | view |
| 8k | 636.7 | 15.8 | 73.3 GB | current |
| 16k | 588.8 | 15.3 | 75.8 GB | view |
| 32k | 481.0 | 14.1 | 80.6 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 17.4 | 1.00× |
| 2× | 31.6 | 1.82× |
| 4× | 52.0 | 2.99× |