← Back to community benchmarks
Qwen3.5-35B-A3B-Hereticmxfp8
Performance
16k
tokens
1,292
PP tok/s
56.7
TG tok/s
12682
TTFT (ms)
35.7
Peak mem (GB)
Hardware
Chip
M3 Max (40c)
Memory
64 GB
GPU Cores
40
Software
oMLX
v0.2.7
macOS
macOS 26.3.1
Context
16,384
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,385 | 69.7 | 34.9 GB | view |
| 4k | 1,440 | 65.8 | 35.0 GB | view |
| 8k | 1,397 | 63.6 | 35.3 GB | view |
| 16k | 1,292 | 56.7 | 35.7 GB | current |
| 32k | 1,118 | 47.8 | 36.5 GB | view |
| 64k | 865.9 | 33.1 | 38.2 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 69.7 | 1.00× |
| 2× | 110.7 | 1.59× |
| 4× | 138.1 | 1.98× |