← Back to community benchmarks
Qwen3.5-27B-Claude-4.6-Opus-Distilled
Performance
8k
tokens
106.8
PP tok/s
13.2
TG tok/s
76730
TTFT (ms)
16.6
Peak mem (GB)
Hardware
Chip
M4 Pro (20c)
Memory
64 GB
GPU Cores
20
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 | 121.5 | 15.0 | 15.8 GB | view |
| 4k | 115.6 | 13.3 | 16.1 GB | view |
| 8k | 106.8 | 13.2 | 16.6 GB | current |
| 16k | 109.4 | 11.5 | 17.5 GB | view |
| 32k | 101.0 | 8.9 | 19.2 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 15.0 | 1.00× |
| 2× | 27.1 | 1.81× |
| 4× | 29.6 | 1.97× |
| 8× | 30.1 | 2.01× |