← Back to community benchmarks
Qwen3.6-27B-oQ4-mtp
Performance
1k
tokens
233.1
PP tok/s
38.4
TG tok/s
4393
TTFT (ms)
16.7
Peak mem (GB)
Hardware
Chip
M4 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.3.9rc1
macOS
macOS 26.4.1
Context
1,024
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 233.1 | 38.4 | 16.7 GB | current |
| 4k | 190.1 | 31.8 | 18.2 GB | view |
| 8k | 177.0 | 30.2 | 19.2 GB | view |
| 16k | 177.4 | 25.8 | 20.7 GB | view |
| 32k | 159.2 | 24.8 | 23.7 GB | view |
| 64k | 145.6 | 20.3 | 29.7 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 38.4 | 1.00× |
| 2× | 36.2 | 0.94× |
| 4× | 41.3 | 1.08× |