← Back to community benchmarks
Qwen3.6-35B-A3B-Kimi-K2.6-Reasoning-Distilled-oQ6e
Performance
8k
tokens
1,903
PP tok/s
72.2
TG tok/s
4305
TTFT (ms)
28.7
Peak mem (GB)
Hardware
Chip
M2 Ultra (60c)
Memory
192 GB
GPU Cores
60
Software
oMLX
v0.3.8
macOS
macOS 26.4.1
Context
8,192
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,296 | 73.7 | 27.6 GB | view |
| 4k | 1,811 | 69.8 | 28.4 GB | view |
| 8k | 1,903 | 72.2 | 28.7 GB | current |
| 16k | 1,840 | 64.4 | 29.6 GB | view |
| 32k | 1,626 | 62.9 | 30.7 GB | view |
| 64k | 1,184 | 53.8 | 34.1 GB | view |
| 128k | 685.7 | 42.4 | 38.8 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 73.7 | 1.00× |
| 2× | 119.0 | 1.61× |
| 4× | 178.5 | 2.42× |
| 8× | 248.3 | 3.37× |