← Back to community benchmarks
Qwen3.5-4B
Performance
64k
tokens
255.4
PP tok/s
16.6
TG tok/s
256562
TTFT (ms)
8.0
Peak mem (GB)
Hardware
Chip
M4 (10c)
Memory
16 GB
GPU Cores
10
Software
oMLX
v0.2.7
macOS
macOS 26.3.1
Context
65,536
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 408.5 | 38.2 | 3.9 GB | view |
| 4k | 405.7 | 36.3 | 4.2 GB | view |
| 8k | 393.0 | 33.9 | 4.6 GB | view |
| 16k | 368.9 | 29.8 | 4.8 GB | view |
| 32k | 320.7 | 23.1 | 5.9 GB | view |
| 64k | 255.4 | 16.6 | 8.0 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 38.2 | 1.00× |
| 2× | 73.2 | 1.92× |
| 4× | 91.3 | 2.39× |
| 8× | 94.1 | 2.46× |