← Back to community benchmarks
Qwen3.5-35B-A3B-OptiQ
Performance
16k
tokens
1,110
PP tok/s
69.0
TG tok/s
14755
TTFT (ms)
23.7
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
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,123 | 78.3 | 21.7 GB | view |
| 4k | 1,204 | 74.9 | 22.5 GB | view |
| 16k | 1,110 | 69.0 | 23.7 GB | current |
| 32k | 998.5 | 60.8 | 25.3 GB | view |
| 64k | 793.4 | 54.6 | 27.8 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 78.3 | 1.00× |
| 2× | 136.3 | 1.74× |
| 4× | 274.8 | 3.51× |