← Back to community benchmarks
Qwen3___5-4B
Performance
64k
tokens
257.2
PP tok/s
16.4
TG tok/s
254817
TTFT (ms)
8.0
Peak mem (GB)
Hardware
Chip
M4 (10c)
Memory
16 GB
GPU Cores
10
Software
oMLX
v0.2.6
macOS
macOS 26.3.1
Context
65,536
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 407.8 | 39.2 | 3.9 GB | view |
| 4k | 404.3 | 37.1 | 4.2 GB | view |
| 8k | 388.7 | 34.5 | 4.6 GB | view |
| 16k | 362.8 | 30.2 | 4.8 GB | view |
| 32k | 320.7 | 23.5 | 5.9 GB | view |
| 64k | 257.2 | 16.4 | 8.0 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 39.2 | 1.00× |
| 2× | 75.4 | 1.92× |
| 4× | 94.2 | 2.40× |