← Back to community benchmarks
Qwen3.5-122B-A10B
Performance
8k
tokens
712.9
PP tok/s
48.2
TG tok/s
11492
TTFT (ms)
67.0
Peak mem (GB)
Hardware
Chip
M3 Ultra (60c)
Memory
256 GB
GPU Cores
60
Software
oMLX
v0.2.6
macOS
macOS 26.3
Context
8,192
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 704.1 | 53.8 | 66.3 GB | view |
| 4k | 729.1 | 51.2 | 66.7 GB | view |
| 8k | 712.9 | 48.2 | 67.0 GB | current |
| 16k | 670.3 | 41.2 | 67.7 GB | view |
| 32k | 590.2 | 35.0 | 69.1 GB | view |
| 64k | 466.9 | 26.0 | 71.9 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 53.8 | 1.00× |
| 2× | 86.0 | 1.60× |
| 4× | 124.1 | 2.31× |
| 8× | 168.6 | 3.13× |