← Back to community benchmarks
Qwen3.5-35B-A3B-OptiQ
Performance
8k
tokens
1,181
PP tok/s
70.2
TG tok/s
6936
TTFT (ms)
22.9
Peak mem (GB)
Hardware
Chip
M2 Ultra (60c)
Memory
64 GB
GPU Cores
60
Software
oMLX
v0.5.1
macOS
macOS 26.5.2
Context
8,192
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,114 | 75.7 | 21.7 GB | view |
| 4k | 1,213 | 73.7 | 22.5 GB | view |
| 8k | 1,181 | 70.2 | 22.9 GB | current |
| 16k | 1,110 | 67.0 | 23.5 GB | view |
| 32k | 997.9 | 59.7 | 25.0 GB | view |
| 64k | 790.7 | 54.3 | 27.9 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 75.7 | 1.00× |
| 2× | 123.1 | 1.63× |
| 4× | 246.4 | 3.25× |
| 8× | 520.5 | 6.88× |