← Back to community benchmarks
Qwen3.6-35B-A3B-oQ8-fp16-mtp
Performance
64k
tokens
465.3
PP tok/s
27.4
TG tok/s
140848
TTFT (ms)
37.8
Peak mem (GB)
Hardware
Chip
M3 Max (30c)
Memory
96 GB
GPU Cores
30
Software
oMLX
v0.5.1
macOS
macOS 26.5.1
Context
65,536
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 682.9 | 44.6 | 35.4 GB | view |
| 4k | 815.2 | 44.1 | 36.1 GB | view |
| 8k | 849.8 | 44.1 | 36.3 GB | view |
| 16k | 813.7 | 43.5 | 36.5 GB | view |
| 32k | 697.6 | 40.9 | 37.0 GB | view |
| 64k | 465.3 | 27.4 | 37.8 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 44.6 | 1.00× |
| 2× | 63.9 | 1.43× |
| 4× | 77.8 | 1.74× |
| 8× | 79.0 | 1.77× |