← Back to community benchmarks
Qwen3.5-35B-A3B-Claude-4.6-Opus-Reasoning-Distilled
Performance
4k
tokens
1,607
PP tok/s
78.0
TG tok/s
2549
TTFT (ms)
36.2
Peak mem (GB)
Hardware
Chip
M4 Max (40c)
Memory
128 GB
GPU Cores
40
Software
oMLX
v0.3.6
macOS
macOS 26.3.1
Context
4,096
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,265 | 80.3 | 35.4 GB | view |
| 4k | 1,607 | 78.0 | 36.2 GB | current |
| 8k | 1,643 | 74.2 | 36.5 GB | view |
| 16k | 1,472 | 66.0 | 37.1 GB | view |
| 32k | 1,268 | 56.3 | 38.5 GB | view |
| 64k | 987.7 | 41.8 | 41.2 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 80.3 | 1.00× |
| 2× | 128.7 | 1.60× |
| 4× | 196.8 | 2.45× |
| 8× | 251.5 | 3.13× |