← Back to community benchmarks
Qwen2.5-VL-3B-Instruct
Performance
1k
tokens
583.7
PP tok/s
15.8
TG tok/s
1754
TTFT (ms)
7.4
Peak mem (GB)
Hardware
Chip
M4 (10c)
Memory
16 GB
GPU Cores
10
Software
oMLX
v0.2.8
macOS
macOS 26.2
Context
1,024
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 583.7 | 15.8 | 7.4 GB | current |
| 4k | 568.7 | 15.4 | 7.6 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 15.8 | 1.00× |
| 2× | 20.7 | 1.31× |
| 4× | 40.7 | 2.58× |