← Back to community benchmarks
Qwen3.5-4B
Performance
16k
tokens
334.9
PP tok/s
18.9
TG tok/s
48917
TTFT (ms)
4.8
Peak mem (GB)
Hardware
Chip
M4 (10c)
Memory
16 GB
GPU Cores
10
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 | 372.6 | 24.8 | 3.9 GB | view |
| 4k | 362.0 | 27.0 | 4.2 GB | view |
| 8k | 356.3 | 25.1 | 4.6 GB | view |
| 16k | 334.9 | 18.9 | 4.8 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 24.8 | 1.00× |
| 2× | 45.5 | 1.83× |
| 4× | 71.2 | 2.87× |