← Back to community benchmarks
Qwen3-4B-Instruct-2507-heretic
Performance
1k
tokens
663.0
PP tok/s
28.4
TG tok/s
1545
TTFT (ms)
8.0
Peak mem (GB)
Hardware
Chip
M4 Pro (16c)
Memory
24 GB
GPU Cores
16
Software
oMLX
v0.3.6
macOS
macOS 26.4.1
Context
1,024
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 663.0 | 28.4 | 8.0 GB | current |
| 4k | 668.8 | 25.9 | 8.6 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 28.4 | 1.00× |
| 2× | 24.1 | 0.85× |
| 4× | 45.5 | 1.60× |