← Back to community benchmarks
Qwen3.6-35B-A3BoQ8
Performance
16k
tokens
2,361
PP tok/s
29.1
TG tok/s
6940
TTFT (ms)
57.0
Peak mem (GB)
Hardware
Chip
M3 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.3.7
macOS
macOS 26.4.1
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 906.1 | 65.0 | 54.8 GB | view |
| 4k | 1,376 | 61.8 | 55.6 GB | view |
| 8k | 1,389 | 60.1 | 55.9 GB | view |
| 16k | 2,361 | 29.1 | 57.0 GB | current |
| 32k | 2,374 | 29.4 | 57.7 GB | view |
| 64k | 2,032 | 28.9 | 59.6 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 65.0 | 1.00× |
| 2× | 109.8 | 1.69× |
| 4× | 157.0 | 2.42× |