← Back to community benchmarks
Qwen3.5-4B
Performance
8k
tokens
386.5
PP tok/s
32.1
TG tok/s
21196
TTFT (ms)
4.6
Peak mem (GB)
Hardware
Chip
M4 (10c)
Memory
16 GB
GPU Cores
10
Software
oMLX
v0.2.6
macOS
macOS 26.3.1
Context
8,192
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 394.7 | 35.4 | 3.9 GB | view |
| 4k | 401.2 | 34.1 | 4.2 GB | view |
| 8k | 386.5 | 32.1 | 4.6 GB | current |
| 16k | 358.0 | 24.4 | 4.8 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 35.4 | 1.00× |
| 2× | 41.1 | 1.16× |
| 4× | 83.9 | 2.37× |
| 8× | 86.8 | 2.45× |