← Back to community benchmarks
Qwen3.5-27B-HereticoQ6
Performance
8k
tokens
168.9
PP tok/s
14.0
TG tok/s
48500
TTFT (ms)
25.4
Peak mem (GB)
Hardware
Chip
M1 Ultra (64c)
Memory
128 GB
GPU Cores
64
Software
oMLX
v0.2.20
macOS
macOS 26.4
Context
8,192
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 169.9 | 21.8 | 23.2 GB | view |
| 4k | 172.0 | 17.9 | 24.8 GB | view |
| 8k | 168.9 | 14.0 | 25.4 GB | current |
| 16k | 167.5 | 2.7 | 26.7 GB | view |
| 32k | 158.9 | 3.0 | 29.2 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 21.8 | 1.00× |
| 2× | 26.2 | 1.20× |
| 4× | 27.6 | 1.27× |