← Back to community benchmarks
Qwen3.5-4B
Performance
1k
tokens
408.8
PP tok/s
37.2
TG tok/s
2505
TTFT (ms)
3.9
Peak mem (GB)
Hardware
Chip
M4 (10c)
Memory
16 GB
GPU Cores
10
Software
oMLX
v0.2.13
macOS
macOS 26.2
Context
1,024
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 37.2 | 1.00× |
| 2× | 73.6 | 1.98× |
| 4× | 94.1 | 2.53× |
| 8× | 93.6 | 2.52× |