← Back to community benchmarks
Qwen3.5-35B-A3B-Claude-4.6-Opus-Reasoning-Distilled
Performance
32k
tokens
562.8
PP tok/s
27.9
TG tok/s
58224
TTFT (ms)
37.0
Peak mem (GB)
Hardware
Chip
M4 Pro (20c)
Memory
64 GB
GPU Cores
20
Software
oMLX
v0.2.10
macOS
macOS 26.2
Context
32,768
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 657.2 | 49.9 | 35.3 GB | view |
| 4k | 778.9 | 48.1 | 35.5 GB | view |
| 8k | 748.6 | 45.6 | 35.7 GB | view |
| 16k | 689.9 | 42.5 | 36.1 GB | view |
| 32k | 562.8 | 27.9 | 37.0 GB | current |
| 64k | 456.2 | 19.6 | 38.6 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 49.9 | 1.00× |
| 2× | 78.5 | 1.57× |
| 4× | 102.2 | 2.05× |
| 8× | 135.5 | 2.72× |