← Back to community benchmarks
Qwen3.5-9B
Performance
16k
tokens
758.8
PP tok/s
62.8
TG tok/s
21591
TTFT (ms)
9.6
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 | 775.8 | 74.3 | 8.7 GB | view |
| 4k | 783.9 | 72.1 | 8.9 GB | view |
| 8k | 776.7 | 68.7 | 9.2 GB | view |
| 16k | 758.8 | 62.8 | 9.6 GB | current |
| 32k | 724.1 | 52.6 | 10.6 GB | view |
| 64k | 653.8 | 39.8 | 12.7 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 74.3 | 1.00× |
| 2× | 103.6 | 1.39× |
| 4× | 122.3 | 1.65× |
| 8× | 137.0 | 1.84× |