← Back to community benchmarks
Qwen2.5-72B-Instruct
Performance
16k
tokens
232.9
PP tok/s
10.2
TG tok/s
70356
TTFT (ms)
46.5
Peak mem (GB)
Hardware
Chip
M5 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.4.4
macOS
macOS 26.5
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 329.4 | 13.1 | 38.9 GB | view |
| 4k | 285.8 | 12.7 | 40.5 GB | view |
| 8k | 251.8 | 11.0 | 42.8 GB | view |
| 16k | 232.9 | 10.2 | 46.5 GB | current |
| 32k | 198.1 | 8.1 | 54.0 GB | view |
| 64k | 154.7 | 6.4 | 69.0 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 13.1 | 1.00× |
| 2× | 17.8 | 1.36× |