← Back to community benchmarks
Qwen3.5-27B
Performance
32k
tokens
464.5
PP tok/s
22.4
TG tok/s
70538
TTFT (ms)
29.7
Peak mem (GB)
Hardware
Chip
M4 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.3.5.dev1
macOS
macOS 26.3.1
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 241.8 | 29.0 | 24.7 GB | view |
| 4k | 250.0 | 27.9 | 26.1 GB | view |
| 8k | 248.1 | 27.0 | 26.8 GB | view |
| 16k | 480.5 | 22.9 | 28.5 GB | view |
| 32k | 464.5 | 22.4 | 29.7 GB | current |
| 64k | 428.4 | 21.1 | 31.6 GB | view |
| 128k | 362.3 | 13.2 | 53.8 GB | view |
| 195k | 307.9 | 18.0 | 70.4 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 29.0 | 1.00× |
| 2× | 54.4 | 1.88× |
| 4× | 62.8 | 2.17× |