← Back to community benchmarks
Qwen3.6-35B-A3B-mxfp8
Performance
32k
tokens
2,733
PP tok/s
81.8
TG tok/s
11992
TTFT (ms)
37.7
Peak mem (GB)
Hardware
Chip
M5 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.5.1
macOS
macOS 26.5.2
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 3,403 | 95.3 | 34.4 GB | view |
| 4k | 3,502 | 93.7 | 35.2 GB | view |
| 16k | 3,369 | 89.0 | 36.1 GB | view |
| 32k | 2,733 | 81.8 | 37.7 GB | current |
| 64k | 1,928 | 68.1 | 40.6 GB | view |
| 128k | 1,152 | 48.3 | 46.1 GB | view |
| 195k | 764.9 | 40.0 | 51.7 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 95.3 | 1.00× |
| 2× | 153.8 | 1.61× |
| 4× | 201.4 | 2.11× |