← Back to community benchmarks
Qwen3.5-9B
Performance
16k
tokens
188.3
PP tok/s
16.2
TG tok/s
87006
TTFT (ms)
7.5
Peak mem (GB)
Hardware
Chip
M4 (10c)
Memory
16 GB
GPU Cores
10
Software
oMLX
v0.2.6
macOS
macOS 26.4
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 216.5 | 18.6 | 6.5 GB | view |
| 4k | 210.5 | 17.7 | 6.8 GB | view |
| 8k | 197.2 | 17.5 | 7.0 GB | view |
| 16k | 188.3 | 16.2 | 7.5 GB | current |
| 32k | 176.1 | 13.2 | 8.5 GB | view |
| 64k | 152.7 | 10.4 | 10.6 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 18.6 | 1.00× |
| 2× | 36.9 | 1.98× |
| 4× | 47.2 | 2.54× |
| 8× | 48.5 | 2.61× |