← Back to community benchmarks
Qwen3.6-35B-A3B-Qwable-Holo3-Qwopus-qx64-hi
Performance
16k
tokens
1,565
PP tok/s
86.9
TG tok/s
10468
TTFT (ms)
31.0
Peak mem (GB)
Hardware
Chip
M4 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.5.1
macOS
macOS 26.3
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,585 | 96.1 | 29.0 GB | view |
| 4k | 1,776 | 93.9 | 29.8 GB | view |
| 8k | 1,717 | 92.1 | 30.3 GB | view |
| 16k | 1,565 | 86.9 | 31.0 GB | current |
| 32k | 1,337 | 80.2 | 32.5 GB | view |
| 64k | 1,034 | 68.3 | 35.4 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 96.1 | 1.00× |
| 2× | 160.4 | 1.67× |
| 4× | 234.1 | 2.44× |
| 8× | 306.6 | 3.19× |