← Back to community benchmarks
Qwen3.5-397B-A17B
Performance
32k
tokens
440.8
PP tok/s
18.5
TG tok/s
74343
TTFT (ms)
213.2
Peak mem (GB)
Hardware
Chip
M3 Ultra (80c)
Memory
512 GB
GPU Cores
80
Software
oMLX
v0.2.10
macOS
macOS 15.7.4
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 481.4 | 35.3 | 210.4 GB | view |
| 4k | 514.1 | 33.9 | 210.7 GB | view |
| 8k | 510.9 | 32.2 | 211.0 GB | view |
| 16k | 489.3 | 25.6 | 211.8 GB | view |
| 32k | 440.8 | 18.5 | 213.2 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 35.3 | 1.00× |
| 2× | 55.8 | 1.58× |
| 4× | 82.4 | 2.33× |
| 8× | 115.5 | 3.27× |