← Back to community benchmarks
Qwen3.5-27B-Claude-4.6-Opus-Distilled
Performance
8k
tokens
223.7
PP tok/s
25.1
TG tok/s
36619
TTFT (ms)
16.6
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
8,192
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 212.7 | 25.9 | 15.8 GB | view |
| 4k | 223.0 | 24.9 | 16.1 GB | view |
| 8k | 223.7 | 25.1 | 16.6 GB | current |
| 16k | 217.2 | 20.4 | 17.5 GB | view |
| 32k | 207.8 | 14.1 | 19.2 GB | view |
| 64k | 186.2 | 9.4 | 22.8 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 25.9 | 1.00× |
| 2× | 48.6 | 1.88× |
| 4× | 58.3 | 2.25× |