← Back to community benchmarks
Qwen3.5-122B-A10B
Performance
16k
tokens
444.0
PP tok/s
36.3
TG tok/s
36903
TTFT (ms)
67.7
Peak mem (GB)
Hardware
Chip
M2 Ultra (76c)
Memory
128 GB
GPU Cores
76
Software
oMLX
v0.2.6
macOS
macOS 26.3
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 439.6 | 47.8 | 66.3 GB | view |
| 4k | 463.3 | 45.6 | 66.7 GB | view |
| 8k | 459.9 | 41.3 | 67.0 GB | view |
| 16k | 444.0 | 36.3 | 67.7 GB | current |
| 32k | 409.4 | 28.3 | 69.1 GB | view |
| 64k | 331.7 | 19.0 | 71.9 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 47.8 | 1.00× |
| 2× | 74.3 | 1.55× |
| 4× | 100.7 | 2.11× |
| 8× | 125.1 | 2.62× |