← Back to community benchmarks
Qwen3.5-397B-A17B
Performance
16k
tokens
384.0
PP tok/s
24.8
TG tok/s
42671
TTFT (ms)
211.8
Peak mem (GB)
Hardware
Chip
M3 Ultra (60c)
Memory
256 GB
GPU Cores
60
Software
oMLX
v0.2.6
macOS
macOS 26.3
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 381.0 | 37.0 | 210.4 GB | view |
| 4k | 404.5 | 35.3 | 210.7 GB | view |
| 8k | 400.7 | 31.3 | 211.0 GB | view |
| 16k | 384.0 | 24.8 | 211.8 GB | current |
| 32k | 350.4 | 17.8 | 213.2 GB | view |
| 64k | 294.3 | 11.7 | 216.2 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 37.0 | 1.00× |
| 2× | 56.6 | 1.53× |
| 4× | 76.1 | 2.06× |
| 8× | 102.1 | 2.76× |