← Back to community benchmarks
Qwen3.5-122B-A10B
Performance
32k
tokens
468.6
PP tok/s
25.3
TG tok/s
69928
TTFT (ms)
72.1
Peak mem (GB)
Hardware
Chip
M4 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.2.22
macOS
macOS 26.3.1
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 567.0 | 47.3 | 66.4 GB | view |
| 4k | 599.6 | 44.8 | 67.9 GB | view |
| 8k | 586.7 | 41.7 | 68.5 GB | view |
| 16k | 544.7 | 36.6 | 69.7 GB | view |
| 32k | 468.6 | 25.3 | 72.1 GB | current |
| 64k | 375.1 | 19.5 | 77.0 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 47.3 | 1.00× |
| 2× | 73.5 | 1.55× |
| 4× | 98.0 | 2.07× |