← Back to community benchmarks
Qwen3.5-122B-A10B
Performance
32k
tokens
325.5
PP tok/s
19.6
TG tok/s
100680
TTFT (ms)
126.0
Peak mem (GB)
Hardware
Chip
M2 Ultra (60c)
Memory
192 GB
GPU Cores
60
Software
oMLX
v0.2.7
macOS
macOS 15.2
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 322.8 | 37.6 | 123.2 GB | view |
| 4k | 361.7 | 35.8 | 123.5 GB | view |
| 8k | 363.5 | 31.7 | 123.9 GB | view |
| 16k | 350.8 | 26.6 | 124.5 GB | view |
| 32k | 325.5 | 19.6 | 126.0 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 37.6 | 1.00× |
| 2× | 59.0 | 1.57× |
| 4× | 78.4 | 2.09× |