← Back to community benchmarks
Qwen3.5-9B
Performance
16k
tokens
416.4
PP tok/s
41.4
TG tok/s
39345
TTFT (ms)
7.6
Peak mem (GB)
Hardware
Chip
M4 Pro (20c)
Memory
64 GB
GPU Cores
20
Software
oMLX
v0.2.6
macOS
macOS 26.3.1
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 443.7 | 49.3 | 6.7 GB | view |
| 4k | 447.6 | 47.4 | 7.0 GB | view |
| 8k | 443.6 | 45.0 | 7.1 GB | view |
| 16k | 416.4 | 41.4 | 7.6 GB | current |
| 32k | 388.4 | 33.6 | 8.7 GB | view |
| 64k | 350.7 | 23.4 | 10.8 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 49.3 | 1.00× |
| 2× | 92.3 | 1.87× |
| 4× | 102.3 | 2.08× |
| 8× | 107.1 | 2.17× |