← Back to community benchmarks
Qwen3.6-27B-oQ8-fp16-mtp
Performance
16k
tokens
271.9
PP tok/s
24.3
TG tok/s
60267
TTFT (ms)
32.8
Peak mem (GB)
Hardware
Chip
M1 Ultra (64c)
Memory
128 GB
GPU Cores
64
Software
oMLX
v0.3.9.dev1
macOS
macOS 26.5
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 256.7 | 26.5 | 28.8 GB | view |
| 4k | 276.4 | 26.8 | 30.3 GB | view |
| 8k | 277.1 | 26.6 | 31.3 GB | view |
| 16k | 271.9 | 24.3 | 32.8 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 26.5 | 1.00× |
| 2× | 34.2 | 1.29× |
| 4× | 41.0 | 1.55× |