← Back to community benchmarks
Qwen3.6-27B-oQ8-fp16-mtp
Performance
8k
tokens
276.8
PP tok/s
26.0
TG tok/s
29600
TTFT (ms)
31.3
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
8,192
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 257.3 | 27.3 | 28.8 GB | view |
| 4k | 277.2 | 26.7 | 30.3 GB | view |
| 8k | 276.8 | 26.0 | 31.3 GB | current |
| 16k | 272.0 | 24.3 | 32.8 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 27.3 | 1.00× |
| 2× | 34.0 | 1.25× |
| 4× | 41.0 | 1.50× |