← Back to community benchmarks
Qwen3.5-9B-DeepSeek-V4-Flash-oQ4-mtp
Performance
32k
tokens
314.8
PP tok/s
37.1
TG tok/s
104093
TTFT (ms)
7.5
Peak mem (GB)
Hardware
Chip
M4 Pro (16c)
Memory
24 GB
GPU Cores
16
Software
oMLX
v0.5.1
macOS
macOS 26.5.2
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 360.9 | 46.6 | 5.9 GB | view |
| 4k | 368.0 | 45.6 | 6.4 GB | view |
| 8k | 363.8 | 44.1 | 6.5 GB | view |
| 16k | 335.3 | 41.4 | 6.8 GB | view |
| 32k | 314.8 | 37.1 | 7.5 GB | current |
| 64k | 279.1 | 29.9 | 8.7 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 46.6 | 1.00× |
| 2× | 95.4 | 2.05× |