← Back to community benchmarks
Qwen3.5-35B-A3B
Performance
64k
tokens
621.6
PP tok/s
20.4
TG tok/s
105432
TTFT (ms)
23.3
Peak mem (GB)
Hardware
Chip
M4 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.2.6
macOS
macOS 26.2
Context
65,536
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,675 | 113.1 | 20.1 GB | view |
| 4k | 1,753 | 100.1 | 20.2 GB | view |
| 8k | 1,536 | 101.0 | 20.4 GB | view |
| 16k | 1,321 | 88.0 | 20.8 GB | view |
| 32k | 1,058 | 58.9 | 21.6 GB | view |
| 64k | 621.6 | 20.4 | 23.3 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 113.1 | 1.00× |
| 2× | 76.3 | 0.67× |
| 4× | 143.1 | 1.27× |
| 8× | 184.4 | 1.63× |