← Back to community benchmarks
Qwen3.6-35B-A3B
Performance
4k
tokens
2,946
PP tok/s
91.0
TG tok/s
1391
TTFT (ms)
20.0
Peak mem (GB)
Hardware
Chip
M3 Ultra (80c)
Memory
512 GB
GPU Cores
80
Software
oMLX
v0.5.1
macOS
macOS 26.5
Context
4,096
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 2,675 | 89.8 | 19.3 GB | view |
| 4k | 2,946 | 91.0 | 20.0 GB | current |
| 8k | 2,791 | 88.8 | 20.2 GB | view |
| 16k | 2,443 | 83.9 | 20.3 GB | view |
| 32k | 2,094 | 77.7 | 21.0 GB | view |
| 64k | 1,622 | 69.9 | 21.8 GB | view |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 89.8 | 1.00× |
| 2× | 171.3 | 1.91× |