← Back to community benchmarks
Qwen3.6-35B-A3B
Performance
128k
tokens
685.6
PP tok/s
21.2
TG tok/s
191169
TTFT (ms)
30.4
Peak mem (GB)
Hardware
Chip
M4 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.3.7
macOS
macOS 26.4.1
Context
131,072
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,296 | 115.4 | 19.3 GB | view |
| 4k | 1,695 | 109.9 | 20.0 GB | view |
| 8k | 1,698 | 102.5 | 20.4 GB | view |
| 16k | 1,591 | 77.0 | 21.0 GB | view |
| 32k | 1,379 | 60.9 | 22.4 GB | view |
| 64k | 1,064 | 41.0 | 25.0 GB | view |
| 128k | 685.6 | 21.2 | 30.4 GB | current |
| 195k | 508.0 | 12.1 | 36.1 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 115.4 | 1.00× |
| 2× | 184.4 | 1.60× |
| 4× | 271.0 | 2.35× |
| 8× | 354.9 | 3.08× |