← Back to community benchmarks
Qwen3.5-122B-A10B6.5bit
Performance
8k
tokens
920.0
PP tok/s
45.2
TG tok/s
8904
TTFT (ms)
96.0
Peak mem (GB)
Hardware
Chip
M3 Ultra (80c)
Memory
256 GB
GPU Cores
80
Software
oMLX
v0.3.4
macOS
macOS 26.4
Context
8,192
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 715.8 | 48.9 | 94.0 GB | view |
| 4k | 913.7 | 45.0 | 95.3 GB | view |
| 8k | 920.0 | 45.2 | 96.0 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 48.9 | 1.00× |
| 2× | 80.0 | 1.64× |
| 4× | 108.1 | 2.21× |
| 8× | 141.0 | 2.88× |