← Back to community benchmarks
Qwen3.5-35B-A3B
Performance
64k
tokens
1,021
PP tok/s
27.5
TG tok/s
64210
TTFT (ms)
23.3
Peak mem (GB)
Hardware
Chip
M4 Max (40c)
Memory
48 GB
GPU Cores
40
Software
oMLX
v0.2.7
macOS
macOS 15.7.3
Context
65,536
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,621 | 115.6 | 20.1 GB | view |
| 4k | 1,742 | 110.9 | 20.2 GB | view |
| 8k | 1,684 | 101.7 | 20.4 GB | view |
| 16k | 1,550 | 80.4 | 20.8 GB | view |
| 32k | 1,312 | 46.8 | 21.6 GB | view |
| 64k | 1,021 | 27.5 | 23.3 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 115.6 | 1.00× |
| 2× | 188.6 | 1.63× |
| 4× | 274.6 | 2.38× |
| 8× | 356.3 | 3.08× |