← Back to community benchmarks
Qwen3.5-27B-9bit
Performance
32k
tokens
283.8
PP tok/s
17.7
TG tok/s
115466
TTFT (ms)
33.3
Peak mem (GB)
Hardware
Chip
M3 Ultra (60c)
Memory
96 GB
GPU Cores
60
Software
oMLX
v0.2.6
macOS
macOS 26.3.1
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 291.0 | 21.4 | 29.9 GB | view |
| 4k | 308.8 | 21.1 | 30.2 GB | view |
| 8k | 309.0 | 20.5 | 30.6 GB | view |
| 16k | 301.1 | 19.4 | 31.5 GB | view |
| 32k | 283.8 | 17.7 | 33.3 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 21.4 | 1.00× |
| 2× | 37.1 | 1.73× |
| 4× | 48.2 | 2.25× |