← Back to community benchmarks
Qwen3.5-122B-A10B
Performance
64k
tokens
584.5
PP tok/s
23.5
TG tok/s
112129
TTFT (ms)
128.7
Peak mem (GB)
Hardware
Chip
M3 Ultra (80c)
Memory
512 GB
GPU Cores
80
Software
oMLX
v0.2.6
macOS
macOS 26.3
Context
65,536
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 855.1 | 40.6 | 123.2 GB | view |
| 4k | 906.9 | 39.1 | 123.5 GB | view |
| 8k | 893.1 | 37.6 | 123.9 GB | view |
| 16k | 843.9 | 35.4 | 124.5 GB | view |
| 32k | 746.4 | 31.2 | 125.9 GB | view |
| 64k | 584.5 | 23.5 | 128.7 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 40.6 | 1.00× |
| 2× | 66.0 | 1.63× |
| 4× | 94.5 | 2.33× |