← Back to community benchmarks
Qwen3.5-122B-A10B-Text-mxfp4
Performance
16k
tokens
393.0
PP tok/s
30.0
TG tok/s
41687
TTFT (ms)
63.3
Peak mem (GB)
Hardware
Chip
M3 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.2.12
macOS
macOS 15.7.4
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 421.2 | 47.1 | 62.0 GB | view |
| 4k | 466.5 | 43.2 | 62.3 GB | view |
| 8k | 441.6 | 37.3 | 62.6 GB | view |
| 16k | 393.0 | 30.0 | 63.3 GB | current |
| 32k | 363.2 | 20.5 | 64.7 GB | view |
| 64k | 268.6 | 12.3 | 67.5 GB | view |
| 128k | 192.8 | 10.9 | 73.0 GB | view |
| 195k | 144.2 | 6.5 | 79.0 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 47.1 | 1.00× |
| 2× | 58.4 | 1.24× |
| 4× | 72.7 | 1.54× |
| 8× | 91.5 | 1.94× |