← Back to community benchmarks
Qwen3.6-27B-oQ4-fp16-mtp
Performance
1k
tokens
338.6
PP tok/s
26.4
TG tok/s
3024
TTFT (ms)
16.7
Peak mem (GB)
Hardware
Chip
M5 Max (40c)
Memory
64 GB
GPU Cores
40
Software
oMLX
v0.3.9.dev2
macOS
macOS 26.4.1
Context
1,024
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 338.6 | 26.4 | 16.7 GB | current |
| 4k | 445.6 | 25.4 | 18.2 GB | view |
| 8k | 468.0 | 25.6 | 19.2 GB | view |
| 16k | 429.4 | 23.4 | 20.7 GB | view |
| 32k | 292.8 | 20.5 | 23.7 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 26.4 | 1.00× |
| 2× | 25.9 | 0.98× |