← Back to community benchmarks
Qwen3.6-35B-A3B
Performance
16k
tokens
418.0
PP tok/s
39.1
TG tok/s
39192
TTFT (ms)
21.0
Peak mem (GB)
Hardware
Chip
M4 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.3.6
macOS
macOS 26.4.1
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 954.9 | 55.8 | 19.3 GB | view |
| 4k | 481.8 | 53.2 | 20.0 GB | view |
| 8k | 481.7 | 40.5 | 20.4 GB | view |
| 16k | 418.0 | 39.1 | 21.0 GB | current |
| 32k | 385.5 | 33.0 | 22.4 GB | view |
| 64k | 267.1 | 20.9 | 25.0 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 55.8 | 1.00× |
| 2× | 47.1 | 0.84× |
| 4× | 68.5 | 1.23× |
| 8× | 109.2 | 1.96× |