← Back to community benchmarks
Qwen3.5-9B-Claude-4.6-HighIQ-INSTRUCT-HERETIC-UNCENSOREDmxfp8
Performance
32k
tokens
717.1
PP tok/s
29.4
TG tok/s
45694
TTFT (ms)
12.2
Peak mem (GB)
Hardware
Chip
M4 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.2.7
macOS
macOS 26.3.1
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 792.1 | 46.7 | 10.2 GB | view |
| 4k | 806.8 | 44.7 | 10.4 GB | view |
| 8k | 777.1 | 41.4 | 10.6 GB | view |
| 16k | 755.6 | 38.7 | 11.2 GB | view |
| 32k | 717.1 | 29.4 | 12.2 GB | current |
| 64k | 626.1 | 20.2 | 14.2 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 46.7 | 1.00× |
| 2× | 68.0 | 1.46× |
| 4× | 78.2 | 1.67× |