← Back to community benchmarks
Qwen2.5-7B-Instruct
Performance
16k
tokens
194.6
PP tok/s
17.3
TG tok/s
84208
TTFT (ms)
5.5
Peak mem (GB)
Hardware
Chip
M4 (10c)
Memory
16 GB
GPU Cores
10
Software
oMLX
v0.2.7
macOS
macOS 15.6.1
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 251.4 | 22.9 | 4.5 GB | view |
| 4k | 243.6 | 21.7 | 4.8 GB | view |
| 8k | 219.2 | 20.3 | 5.0 GB | view |
| 16k | 194.6 | 17.3 | 5.5 GB | current |
| 32k | 164.8 | 12.9 | 6.4 GB | view |
| 64k | 120.1 | 8.3 | 8.3 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 22.9 | 1.00× |
| 2× | 42.4 | 1.85× |
| 4× | 59.6 | 2.60× |
| 8× | 60.3 | 2.63× |