← Back to community benchmarks
Qwen3.5-122B-A10B
Performance
16k
tokens
1,432
PP tok/s
52.5
TG tok/s
11439
TTFT (ms)
68.9
Peak mem (GB)
Hardware
Chip
M5 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.3.2
macOS
macOS 26.4
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,063 | 66.0 | 65.6 GB | view |
| 4k | 1,503 | 62.7 | 67.0 GB | view |
| 8k | 1,544 | 59.9 | 67.7 GB | view |
| 16k | 1,432 | 52.5 | 68.9 GB | current |
| 32k | 1,161 | 42.9 | 71.2 GB | view |
| 64k | 872.4 | 30.6 | 76.1 GB | view |
| 128k | 525.0 | 19.4 | 85.7 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 66.0 | 1.00× |
| 2× | 95.8 | 1.45× |
| 4× | 118.0 | 1.79× |
| 8× | 141.2 | 2.14× |