← Back to community benchmarks
Qwen3.5-35B-A3B
Performance
64k
tokens
940.4
PP tok/s
25.1
TG tok/s
69687
TTFT (ms)
39.5
Peak mem (GB)
Hardware
Chip
M2 Ultra (76c)
Memory
192 GB
GPU Cores
76
Software
oMLX
v0.2.6
macOS
macOS 15.0
Context
65,536
Performance by Context Length
| Context | PP tok/s | TG tok/s | Peak Mem | |
|---|---|---|---|---|
| 1k | 1,253 | 56.8 | 36.2 GB | view |
| 4k | 1,371 | 56.7 | 36.3 GB | view |
| 8k | 1,354 | 45.1 | 36.5 GB | view |
| 16k | 1,293 | 46.3 | 37.0 GB | view |
| 32k | 1,167 | 34.0 | 37.8 GB | view |
| 64k | 940.4 | 25.1 | 39.5 GB | current |
Batching Results
| Batch Size | TG tok/s | Speedup |
|---|---|---|
| 1× | 56.8 | 1.00× |
| 2× | 86.4 | 1.52× |
| 4× | 161.8 | 2.85× |
| 8× | 272.5 | 4.80× |