← Back to community benchmarks
Qwen3.5-27B-Claude-4.6-Opus-Distilled
Performance
8k
tokens
28.6
PP tok/s
2.8
TG tok/s
286846
TTFT (ms)
16.8
Peak mem (GB)
Hardware
Chip
M4 (10c)
Memory
32 GB
GPU Cores
10
Software
oMLX
v0.3.12
macOS
macOS 26.2
Context
8,192
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 54.6 | 6.6 | 15.8 GB | view |
| 4k | 45.9 | 3.4 | 16.9 GB | view |
| 8k | 28.6 | 2.8 | 16.8 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 6.6 | 1.00× |
| 2× | 5.6 | 0.85× |
| 4× | 7.0 | 1.06× |