← Back to community benchmarks
Qwen3.5-27B-Claude-4.6-Opus-Distilled
Performance
4k
tokens
163.8
PP tok/s
16.8
TG tok/s
25006
TTFT (ms)
16.1
Peak mem (GB)
Hardware
Chip
M3 Max (30c)
Memory
36 GB
GPU Cores
30
Software
oMLX
v0.2.11
macOS
macOS 26.3
Context
4,096
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 158.5 | 17.8 | 15.8 GB | view |
| 4k | 163.8 | 16.8 | 16.1 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 17.8 | 1.00× |
| 2× | 33.7 | 1.89× |
| 4× | 40.7 | 2.29× |