← Back to community benchmarks
Qwen3.5-9B-Claude-4.6-Opus-Reasoning-Distilled
Performance
8k
tokens
208.6
PP tok/s
20.3
TG tok/s
39267
TTFT (ms)
6.2
Peak mem (GB)
Hardware
Chip
M4 (10c)
Memory
32 GB
GPU Cores
10
Software
oMLX
v0.2.7
macOS
macOS 26.2
Context
8,192
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 213.4 | 20.3 | 5.7 GB | view |
| 4k | 221.5 | 20.9 | 5.9 GB | view |
| 8k | 208.6 | 20.3 | 6.2 GB | current |
| 16k | 200.1 | 18.7 | 6.7 GB | view |
| 32k | 186.3 | 16.2 | 7.7 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 20.3 | 1.00× |
| 2× | 42.7 | 2.10× |
| 4× | 50.9 | 2.51× |
| 8× | 52.0 | 2.56× |