← Back to community benchmarks
Qwen3.5-122B-A10B
Performance
16k
tokens
331.8
PP tok/s
24.9
TG tok/s
49380
TTFT (ms)
96.1
Peak mem (GB)
Hardware
Chip
M1 Ultra (64c)
Memory
128 GB
GPU Cores
64
Software
oMLX
v0.2.14
macOS
macOS 26.3
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 315.0 | 32.3 | 94.8 GB | view |
| 4k | 341.5 | 31.2 | 95.1 GB | view |
| 8k | 340.8 | 28.9 | 95.4 GB | view |
| 16k | 331.8 | 24.9 | 96.1 GB | current |
| 32k | 307.3 | 19.5 | 97.5 GB | view |
| 64k | 241.8 | 13.2 | 100.3 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 32.3 | 1.00× |
| 2× | 48.3 | 1.50× |
| 4× | 62.6 | 1.94× |